Documente Academic
Documente Profesional
Documente Cultură
INTERNATIONAL ASSOCIATION OF HYDROGEOLOGISTS SELECTED PAPERS INTERNATIONAL ASSOCIATION OF HYDROGEOLOGISTS SELECTED PAPERS
23
Editors:
Karstified rocks, of various lithologies, cover more than 10% of the land surface of
Neno Kukuric
Neven Kreic
Zoran Stevanovic
Editors:
an informa business
Karst without Boundaries
Selected papers on hydrogeology
23
Series Editor: Dr. Nick S. Robins
Editor-in-Chief IAH Book Series, British Geological Survey, Wallingford, UK
Editors
Zoran Stevanovic
Faculty of Mining and Geology, University of Belgrade,
Belgrade, Serbia
Neven Kreic
Environment and Infrastructure, Amec Foster Wheeler,
Kennesaw, GA, USA
Neno Kukuric
IGRAC, Delft, The Netherlands
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business
2016 Taylor & Francis Group, London, UK
Typeset by Quick Sort (India) Pvt Ltd., Chennai, India
Printed and Bound by CPI Group (UK) Ltd, Croydon, CR0 4YY
All rights reserved. No part of this publication or the information contained
herein may be reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, by photocopying, recording or
otherwise, without prior permission in writing from the publisher. Innovations
reported here may not be used without the approval of the authors.
Although all care is taken to ensure integrity and the quality of this publication
and the information herein, no responsibility is assumed by the publishers nor
the author for any damage to the property or persons as a result of operation
or use of this publication and/or the information contained herein.
Published by:CRC Press/Balkema
P.O. Box 11320, 2301 EH Leiden, The Netherlands
e-mail: Pub.NL@taylorandfrancis.com
www.crcpress.com www.taylorandfrancis.com
Library of Congress Cataloging in Publication Data
Applied for
Preface ix
PART I
Management of transboundary karst aquifers 1
1 Dinaric Karst Aquifer One of the worlds largest transboundary
systems and an ideal location for applying innovative and integrated
water management 3
Z. Stevanovic, N. Kukuric, . Peka, B. Jolovic, A. Pambuku & D. Radojevic
PART 2
Karst aquifer characterisation and monitoring 101
8 Optimal water management Prerequisite for regional socio-economic
development in the karst of the south-eastern Dinarides 103
P. Milanovic
PART 3
The water flow in karst: From vadose to discharge zone 225
16 Recharge processes of karst massifs in southern Italy 227
F. Fiorillo & M. Pagnozzi
Table of contents vii
PART 4
Engineering, sustainable use and protection of water in karst 285
20 Creating environmental impact indicators in dynamic karst
system Dinaric karst case example 287
Z. Stevanovic
Karst covers more than 13% of the continental ice-free surface of our planet, with
major coverage in the Middle East and Central Asia where it occupies approximately
23% of the land surface. According to some estimates more than 20% of the worlds
population is consuming groundwater originated from karst aquifers. At the same
time, the management of karst aquifers and their water resources is more problematic
compared to any other aquifer type due to uncertainties in defining their boundaries,
often irregular drainage regime, and vulnerability to pollution. All these problems
are closely related to high heterogeneity of karst aquifers and their locally very high
permeability.
During recent decades the first systematic overview and inventory of transbound-
ary aquifers and their internationally shared water resources has been made. This work
led by UN agencies (UNESCO, UNECE, IGRAC) and the International Association of
Hydrogeologists (IAH) underlined the challenges of international groundwater data
availability and highlighted the importance of establishing proper monitoring systems
as a prerequisite to the better assessment of shared water resources.
When it comes to karst transboundary aquifer systems there is a rough approx-
imation that not less than 20% of all internationally shared aquifers belong to karst
types. This estimate is currently being adjusted upwards as a result of new investiga-
tions especially in Central and South Eastern Europe, and the Middle East within the
Alpine orogenic belt and its branches, along with the numerous middle- and small-size
countries and their dense network of borders.
A typical example of these complexities is the Dinaric mountain system, the region
of classical karst, which is composed almost entirely of carbonate rocks (limestones
and dolomites), with thickness often exceeding 1000m. In this part of Europe not
only was the term karst born (the German derivation of the local name of the district
between Italy and Slovenia called Carso or Kras), but also a new scientific discipline
karstology was established at the end of the ninetenth century. The appearance of sev-
eral new sovereign states in the 1990s, from what was once Yugoslavia, and complex
transboundary inter-linkages have had a considerable impact on water use and water
sharing for domestic supply, power generation, and agriculture. For instance, this is
also an area where one of the worlds largest springs, located in one country, has 95%
of its catchment in another country. Similarly, this is an area where one may enter a
cave in one country and several kilometers away enter another country. These, among
many others, are the reasons why the DIKTAS (Protection and Sustainable Use of
the Dinaric Karst Transboundary Aquifer System) project was initiated and included
xPreface
three former Yugoslav countries (Croatia, Bosnia & Herzegovina and Montenegro)
and Albania. The project was conducted in the period of 20112015, as full-size GEF
regional project, implemented by UNDP and executed by UNESCO-IHP.
The task of organising the scientific conference related to karst and transboundary
issues was part of the DIKTAS. After four years of work the team of international and
national consultants organised the international conference Karst without Boundaries
which took place on 1115 June of 2014 in Trebinje, Bosnia & Herzegovina, with a
mid-conference field seminar in Dubrovnik, Croatia, and a post-conference excursion
in Boka Kotorska Bay, Montenegro. The conference was preceded by an interna-
tional summer school and a field seminar Characterization and Engineering of Karst
Aquifers supported by UNESCO and attended by university students from Europe,
Asia, and South America. The international summer school is now traditionally
organised beyond DIKTAS as one of its capacity building activities.
This book contains 23 selected chapters from the conference Karst without
Boundaries. These are written by 62 authors coming from 19 countries. The book is
divided into four main topics: Management of transboundary karst aquifers, Karst
aquifer charaterisation and monitoring, Water flow in karst: from vadose to discharge
zone, and Engineering, sustainable use and protection of water in karst.
The chapter Management of transboundary karst aquifers contains seven con-
tributions. The first one aims to improve understanding of transboundary ground
water resources of the Dinaric Karst region and to present DIKTAS project goals and
achievements. The second opens a general but very sensitive question of relevancy
of boundaries in karst. The third presents another important international project
for worldwide karst aquifers mapping (WOKAM). The remaining four contributions
present concerns and results of transboundary aquifers projects between Syria and
Lebanon, Slovakia and Hungary, Bulgaria and Serbia, and Montenegro and Albania.
The chapter Karst aquifer characterisation and monitoring is the largest in this
book with eight contributions. This is also result of common understanding of hydro-
geologists that characterisation and monitoring are essential tools for determining
aquifer behaviour and its resources before any instruments of water policy are to be
imposed. The introductory paper of this group again deals with Dinaric karst and
achievements of implemented large engineering projects which became the basis for
socio-economic development of this region. Three papers of this group deal with aqui-
fer recharge processes, spring hydrographs and time series analyses with examples
from Italy, Slovenia and Iran. Characterisation of aquifer based on hydrochemistry
and several other applied field methods is discussed with case studies from Spain
(evaporitic aquifer) and Hungary (shallow aquifer). Complex karst aquifer charac-
terisation including development of physical models is presented with examples from
Croatia (Plitvice) and Serbia (Beljanica Mountain).
The Water flow in karst: from vadose to discharge zone contains case studies
from Italy, south China, eastern Serbia and Croatia. All four are different in content,
dealing with aquifers in orogenic belt, tower karst, fresh and thermal waters, and
catchment delineation, but all present suggestions how to establish conceptual models
applicable to local conditions.
The final chapter Engineering, sustainable use and protection of water in karst
also contains thematically various subjects. Two examples are related to surface and
groundwater damming in karst and related concerns (examples from China and
Prefacexi
Zoran Stevanovic
Neven Kreic
Neno Kukuric
This page intentionally left blank
Part 1
Management of
transboundary karst aquifers
This page intentionally left blank
Chapter 1
ABSTRACT
The Dinaric region is a karst holotype with its north west margin in the Crasso area around
Trieste in Italy, and south west margin in Albania. There are more than 150 poljes, and in cer-
tain areas the density of the dolines can reach 150/km2. Not only was the term karst born in the
area, but Jovan Cvijic and his followers founded here a new scientific discipline karstology,
at the end of the nineteenth century. The Dinaric region is by far the richest in all of Europe:
there are more than 100 springs with a minimum discharge over 500 l/s. However, there are
numerous problems for groundwater sustainable utilisation and protection from pollution.
Therise of several new sovereign states from what was once Yugoslavia has established com-
plex transboundary inter-linkages that impact water use and water sharing for domestic supply,
power generation, and agriculture. DIKTAS (Protection and Sustainable Use of the Dinaric
Karst Transboundary Aquifer System) is a GEF project implemented by UNDP and UNESCOs
IHP. Its mandate is to improve understanding of shared water resources and to facilitate their
equitable and sustainable utilisation, including the protection of dependent ecosystems.
1.1 INTRODUCTION
The Dinaric system is a karst holotype and a classic karst region. Not only was the
term karst born in the area (the German derivation of the local name of the district
around Trieste between Italy and Slovenia Carso or Kras) but also at the end
of the nineteenth century a group of researchers including A. Penck, A. Grund, and
F. Katzer led by J. Cvijic established a new scientific discipline karstology. Cvijic
undertook most of his work in the Dinaric karst, and in Das Karstphnomen (1893),
argued that rock dissolution was the key process in the creation of most types of
dolines, the diagnostic karst landforms. The Dinaric karst thus becomes the locus
typicus area for dissolutional landforms (Ford, 2005) and some local terms were
accepted, and are still used, in international karst terminology (e.g. ponor, doline,
uvala, polje) (Stevanovic & Mijatovic, 2005). Cvijic stated that there is no deeper
and more thorough karst development than HerzegovinaMontenegros karst located
between the lower Neretva River, Skadar Lake and the Adriatic Sea.
Following Cvijics research, a large number of specialists from former Yugoslavia,
Italy and Albania further improved the knowledge of the Dinarides in terms of
hydrology, geomorphology, geology, hydrogeology and social/humanistic sciences.
Dinaric Karst Aquifer One of the worlds largest transboundary systems 5
Austria
Hungary
Slovenia N
46,000000
46,000000
Romania
Rijeka
Una Tuzla
Zenica
Serbia
Zadar
Krka
Bosnia and Herzegovina
44,000000
44,000000
Sarajevo
Cetina
Split Foca
Mostar
Neretva
Piva
Bilecko
Trebisnjica Lake Niksic
Montenegro
Dubrovnik
Podgorica
(Kosovo)
Cijevna
Budva
Legend Skoder
42,000000
42,000000
State Boundaries
official
Tirana
F.Y.R. Macedonia
Dures
unofficial
Diktas Boundary
Albania
Fijer Berat
TB Aquifers
Korca
TB Aquifers
Towns (population)
1500050000
Greece
40,000000
40,000000
50001100000
100001250000
250001500000
Figure 1.1 Extension of Dinaric karst from Italy to Albania and the boundaries of study area of four
DIKTAS project countries. Some selected aquifers of major transboundary concern are
also shown.
Preparation of the Basic Geological Map of Yugoslavia (some fifty years ago) on the
scale 1:100000 (with working sheets 1:25000) substantially improved the geologi
cal knowledge on Dinaric karst. During several decades, large scale hydrogeologi
cal exploration was carried out for the construction of dams in the Dinaric karst
region. Technical applications of control and regulation of karst aquifers through
the construction of galleries, batteries of wells, and groundwater reservoirs (storage)
represent an important contribution to the international hydrogeological science.
Petar Milanovics Hydrogeology of karst and methods of investigation (first edition
1979) became one of the most important references when dealing with karst ground
water distribution and circulation. A monograph Hydrogeology of the Dinaric Karst
(Mijatovic, 1984) published by IAH, also confirms the wide interest of the hydrogeo-
logical community in the Dinaric karst.
The Dinaric karst is almost entirely carbonate (limestones and dolomites), its
thickness is often over 1000 m and it is mostly of Mesozoic age (Tethys sedimen-
tary basin). The development of the Dinaric karst was gradual. Herak (1972) stated
6 Karst without Boundaries
that at the end of the Triassic or during the Lower Jurassic, the Triassic carbonate
rocks were first exposed to the impact of water circulation processes. At the end of
the Upper Cretaceous and during the Paleocene, intensive uplifting and folding took
place, during which most of the carbonate and flysch rocks were tectonised. After
the Laramian tectonic phase, the next intensive movements occurred in the Helvetian
phase (Eocene/Oligocene). All the main nappes along the Adriatic/Ionian Sea coastline
can be related to this stage and the rising of large land masses, accompanied locally
by intensive structural changes providing the potential for more intensified subsurface
water circulation and widespread karstification (Herak, 1972). Since the Oligocene,
the Dinaric region has been continuously exposed to weathering, providing favor-
able conditions for intensive subsurface water circulation and the development of
karst features. The most distinctive effects can be found in the area of uplifting and
subsidence. The areas of subsidence include the poljes where the water was active
before and after the diastrophic movements. The Pleistocene started not only with
climatic changes (glacial process, lowering sea level) but also brought a new structural
and morphologic evolution. Climate change and the rate of diastrophic movements
regulated the periods of accumulation and later the removal of young deposits from
the poljes (Mijatovic, 1983).
Three major tectonic units are usually distinguished in the Dinarides: External,
Central and Inner Dinarides. These can be additionally separated into several sub-units
(Herak, 1972) (Figure 1.2). Accordingly, the hydrogeological classification of Dinaric
karst indicates the units: Adriatic karst belt, High karst belt, Fluviokarst and Isolated
karst (arin, 1983; Mijatovic, 1984). Although no precise equivalency between tec-
tonic and hydrogeology units exists, the Adriatic belt could be considered an equiv-
alent to External Dinarides, and the High karst to Central, while the last two are
distributed over Inner Dinaric karst.
The Albanian tectonic classification is different in names and in structures.
Nevertheless, the units such as the Adriatic depression (Figure 1.3) and the parts of
the Ionian, Krasta-Cukali or Kruja zones, belong to the External Dinarides, while the
Mirdita unit could be interpreted as an extension of the Central ophiolitic zone of the
Central Dinarides (Mee & Aliaj, 2000).
Dinaric karst is a mountainous region with a prevalence of highly karstified
rocks and large karstic poljes and valleys created in tectonic depressions by peren-
nial or sinking streams. The karstification base is deep; boreholes have locally
encountered karstified zones at depths of 2000 m but Milanovic (2005) con-
cluded that the average depth does not exceed 350400 m. The karstification in
the near-surface zone (010 m) is about 30 times larger than at a depth of 300 m
(Milanovic, 2000). The Dinaric region contains all types of karst landforms and
features including karren (lapies), d olines, jamas (pits), ponors (swallow holes,
sinks), dry and blind valleys, caves and caverns as single forms, and uvalas (Figure
1.4), poljes and karst plains as larger complex forms (Cvijic, 1893; Boicevic,
1966; Roglic, 1972; Kreic, 1988). Asan example, the number of sinkholes (dol-
ines) in certain areas can reach 150 km2. According to Milanovic (2000), in the
Dinaric karst region there are approximately 130 poljes. The total area of all these
poljes is about 1350 km2.
Livanjsko Polje, considered the worlds largest karst polje, covers an area of
380km2, and together with Buko Blato, which morphologically is its integral part, it
Dinaric Karst Aquifer One of the worlds largest transboundary systems 7
0 100 200 km
LJUBJANA
ZAGREB
BEOGRAD
SARAJEVO
AD
RI
AT
IC
SE
A
TITOGRAD
SKOPJE
1 2
3 4 5 6
Figure 1.2 Major tectonic units of Yugoslavia (Herak, 1972). 1 = Outer Dinaric units (Adriatic and
high karst); 2 = Inner Dinaric and south Alpine units; 3 = Pannonian Basin; 4 = eastern Alps;
5 = Serbian-Macedonian Belt; 6 = Carpathian-Balkanides Belt.
totals 433 km2. The poljes are characterised by complicated hydrogeological relations:
drainage of surface water is achieved through many ponors. These are frequently
located in the polje areas nearest to the prevailing erosion base. In the Nikicko polje,
about 880 ponors and estavelles were identified, 851 of which are located along its
southern perimeter. In Popovo Polje there are more than 500 ponors and estavelles.
Some of them are lakes or swamps, the others periodically inundated or even dry. In
general, the poljes are geologically heterogeneous, but an abundance of impervious
rocks implicate tectonic origins.
Herak (1972) stated that more than 12 000 caves have been explored in former
Yugoslavia alone, more than 5000 of which are in Croatia. Kreic (1988) listed some
15 potholes (pits, shafts) in former Yugoslavia deeper than 400 m. In the meantime,
some much deeper potholes have been discovered: Lukina jama-Trojama (1392 m)
and the Slovacka jama (1301 m) at Velebit Mountain, Croatia are among the deep-
est speleological phenomena in the World. At the Kameno more (Stone Sea) and the
Orjen Mountain above Kotor Bay (Montenegro), within an area of only 8 km2, more
8 Karst without Boundaries
Figure 1.3 The contact zone between large tectonic zones Adriatic and Ionian (Dhermi, Albanian
coast, photo Z. Stevanovic).
than 300 vertical shafts were registered (Milanovic, 2005). Some of those shafts were
speleologicaly investigated to depths of 200350 m.
As a result of intensive karstification a network of highly permeable underground
channels acts as preferential pathways of intensive groundwater circulation. Along
with its richness in various karstic features, the Dinaric region is by far the richest in
Europe in water resources but these are unequally distributed throughout the year,
which results from climate and a high degree of karstification (Bonacci, 1987). Some
areas, such as southern Montenegro, are characterised by an intensive water balance
where the average specific yield is over 40 l/s/km2 (Stevanovic et al., 2012). In the
Dinaric region of former Yugoslavia there are 230 springs with a minimum discharge
Dinaric Karst Aquifer One of the worlds largest transboundary systems 9
Figure 1.4 A small depression uvala used for crop cultivation (Orjen Mountain, shared between
Montenegro and Bosnia & Herzegovina, photo Z. Stevanovic).
of over 100 l/s, while about 100 springs have a minimal discharge of over 500 l/s
(Komatina, 1983). In the Albanian karst there are roughly 110 springs with an aver-
age discharge exceeding 100 l/s (Eftimi, 2010).
There are some distinctive patterns in distribution of karstic features in the main
geo-structures of the Dinaric system. In the Adriatic zone, vruljas (submarine springs),
(Milanovic, 2007), are the most specific hydrogeologic features. They are formed by
the sinking of the floor of the northern Adriatic Sea and rising of the sea level, so that
karstified land was submerged beneath the sea (Figure 1.5).
The surface and groundwater of the Dinaric karst belongs to two main catch-
ments. The area of External (Outer) Dinarides belongs to the Adriatic catchment
10 Karst without Boundaries
NW
Ponor Ponor
(sinkhole) Spring (sinkhole) Polje Spring (m)
1000
800
400
200
S. I. 5 0 5 10 km
0
1 2 3 4 5 6 7 8
Figure 1.5 Schematic cross-section through the karstic poljes of the Dinarides (after Mijatovic 1983,
modified by Stevanovic). Legend: 1. karstified Mesozoic limestone; 2. flysch barrier; 3. porous
aquifer of polje; 4. fault; 5. groundwater flow; 6. direction of flow around the barrier;
7. regional flow; 8. groundwater table.
(asmall part to the Ionian Sea), while the Internal (Inner) Dinarides are part of the
Sava (i.e. Danube and the Black Sea catchment area). The main river basins in the
Adriatic/Ionian catchment area are the Vjosa, Seman, Drini, Buna (Bojana), Zeta,
Neretva, Cetina, Krka, Zrmanja and Soca. Karstic groundwater from the river basins
of the Tara, Piva, Vrbas, Pliva, Sana, Una, the upper course of the Kupa River and the
Krka (in Slovenia) gravitates to the Black Sea catchment area (Stevanovic et al., 2012).
Komatina (1983) noted that tracer experiments were conducted at more than
650sites in the Dinaric karst of former Yugoslavia. In eastern Herzegovina alone,
tracers were applied at 281 sites, in the catchment area of the Cetina River at 99 sites
and in the Skadarsko Lake catchment area at 77 sites.
According to Milanovic (2000) more than several hundred investigations have
been performed in the Dinaric karst so far, for the purpose of finding the major routes
of underground water circulation (Figure 1.6). Based on experiments, the average
flow velocity is estimated to vary between a range of 0.00255.2 cm/s. Such extreme
values are rare, whereas an average velocity is about 5 cm/s. Based on 380 conducted
experiments Komatina (1983) concluded that the frequency of groundwater velocities
in Dinaric karst is: in 70% of cases from 0 to 5 cm/s; in 20% of cases 5 to 10 cm/s;
and in 10% of cases more than 10 cm/s. During the dry season and low aquifer water
table, water circulation in the karst system is characterised by slower rates of flow.
The water labeled with dye takes two- to five-fold less time to travel the same distance
during a season of high hydrologic activity (Milanovic, 2000). The same author pres-
ents an example: to cover the distance (34 km) from Gatacko Polje to the Trebinjica
Spring (Dinaric karst, Herzegovina), the underground flow takes 35 days when the
water table is low and inflow is small. During high water levels and large inflow, the
water takes only five days to cover the same distance.
The karst aquifer recharges from precipitation and waters percolating from numer-
ous sinking rivers. Depending on the locality, morphology and karstification properties
the average infiltration rate lies between 50% to 80% of the precipitation. In the Cetina
Dinaric Karst Aquifer One of the worlds largest transboundary systems 11
Figure 1.6 Tracing test at the Ponikve Ponor (Dabarsko Polje, East Herzegovina, photo courtesy of
eljko Zubac).
River catchment area, more than 80% of the precipitation appears at the terminal water
gauge controlled profile, while in the Trebinjica River catchment area this percentage
reaches 90%. Effective infiltration in the Albanian karst is on average made up of 4055%
rainfall (Eftimi, 2010). Considering that the rainfall rate in the region is one of the
highest in Europe (over 2000 mm annually in the southern part, and even up to 5000 mm
in Boka Kotorska Bay) there are abundant but variable water reserves. However, more
than 70% of the precipitation occurred during the wet season (OctoberMarch).
Milanovic (2005) noted that only through three huge springs along the Neretva
Valley and Adriatic coast (Buna, Bunica and Ombla) and a few spring zones in the
Kotor Bay, more than 150 m3/s is discharged annually into the Adriatic Sea directly or
indirectly through the Neretva River.
Three capital cities in the project countries receive drinking water from the karstic
aquifers. Sarajevo obtains a part of its water supply from the Vrelo Bosne springs in
12 Karst without Boundaries
Central Dinarides (1.424 m3/s). The Albanian capital Tirana is supplied in part from
the spring that discharges the Triassic and Jurassic karstic aquifer of Mali me Gropa
plateau (Eftimi, 2010), while downstream from the third important spring, Buvilla,
issuing from Dajti Mountain a large reservoir has been constructed. Mareza spring
(2.010.0 m3/s) in the Skadar basin is the main source of water for the Montenegrin
capital city of Podgorica (Radulovic, 2000). Along the Adriatic, Ionian and Aegean
coast, almost all cities and tourist centres use karstic groundwater (Stevanovic, 2010;
Stevanovic & Eftimi, 2010).
Perhaps the most famous and the largest spring on the northern Italian coast is
Timavo, with an average discharge rate of 30 m3/s. Jadro Spring is the main source
for the water supply of Split (Figure 1.7). The average minimum discharge of Jadro
during the recession period is 35 m3/s, while maximum discharge is often over
50m3/s (Bonacci, 1987).
Figure 1.7 One of the largest springs on the Adriatic coast, Jadro Spring supplies drinking water to Split
(Croatia) which is the main reason why the Romans founded the city (photo Z. Stevanovic).
Dinaric Karst Aquifer One of the worlds largest transboundary systems 13
Figure 1.8 Ombla spring (Rijeka Dubrovacka source, Croatia, photo Z. Stevanovic).
Ombla Spring is the largest permanent karstic spring at the South Adriatic coast
(Figure 1.8). It supplies the city of Dubrovnik; at a maximum discharge of about
154m3/s. Since the completion of the Trebinjica Hydropower System and the regu-
lation of this longest sinking river in Europe in the catchment of Ombla, the average
discharge of Ombla was reduced from 34 m3/s to 24 m3/s. However, the minimum
discharge (2.3 m3/s) is not affected (Milanovic, 2006).
The main springs along Boka Kotorska Bay in Montenegro are: Gurdic and
kurda spring near Kotor, Ljuta spring at Orahovac, Spila spring at Risan, Morinj
springs, Opacica at Herceg Novi and Plavda at Tivat. The Sopot near Risan is a well
known submarine spring (Figure 1.9). All these springs are characterised by a high
variation in discharges due to a highly karstified catchment and extremely fast prop-
agation of the rainfall. Some of those springs even dry up completely during summer
while after intensive rainfall some of them can discharge over 100 m3/s.
In the Skadar Lake basin there is a large number of sublacustrine springs such as
Oko Matice, Golac, Kaludjerovo Oko and many other along the edge of Malo Blato
(Radulovic, 2000). Volac, Karuc, Bolje sestre was recently tapped for the regional
water supply of the Montenegrin coastal area (Stevanovic, 2010).
It is assumed that 2/3 of the whole groundwater resources in Albania are within
karstic aquifers, providing more than 60% of the water consumed in the country
(Eftimi, 2010). The average potential yield of coastal karst aquifers in the catch-
ment of the Ionian Sea in Albania is estimated at about 1520 m3/s (Eftimi, personal
communication).
14 Karst without Boundaries
Figure 1.9 Sopot spring near Risan (Montenegro) during peak flow discharge (left, photo Z. Stevanovic)
and during drought when only submarine flow exists (right, photo S. Milanovic).
A large fluctuation in the water table is common in the region (Peka et al., 2012). For
instance, the water level can change by 312 m during a period of 183 days at the obser-
vation borehole Z-3 in the Nevesinjsko Polje. In the Cetina River basin the maximum
recorded water table increase was 3.17 m/h. Along with continual seasonal flooding
of cultivated land, this was also the main reason why large projects to regulate river
flows were initiated in all the countries in the region after World War Two and many
of these were implemented during the 1960s and 1970s. The idea to regulate flows is
much older; the Klinje Dam (the Munica stream, Boznia Herzegovina) was built in the
period 18881896, while the hydro-electrical power plant (HPP) at Kraljevac (Cetina,
Croatia) was erected in 1912. Today many streams are dammed and their waters are
utilised by HPPs. The major dams and reservoirs have been built on the Cetina, Neretva,
Trebinjica, Zeta, and Drini rivers. Dinaric karst becomes, therefore, a reference area
for the successful completion of dams in karst, a problematic media for water losses
(Milanovic, 2000; 2006). About 2/3 of the total existing hydro power facilities of the
four countries are located in the DIKTAS karst area, therefore, hydro power generation
from the DIKTAS karst system plays significant role in these countrys economies.
The Cetina water system is managed by the Croatian water authority although
a considerable amount of water originates in the territory of Bosnia & Herzegovina
including the catchments of the Kupres, Glamoc, Duvno, and Livno poljes, and Buko
Blato (Figure 1.10). Currently there are five HPPs on the Cetina River: the largest
Dinaric Karst Aquifer One of the worlds largest transboundary systems 15
Figure 1.10 Lake Cetina, upstream from the dam (photo Z. Stevanovic).
s torage reservoir, Buko Blato, has a capacity of 831 106 m3. Bonacci (1987) stated
that Cetinas surface area is from 3700 to 4300 km2 of which the topographic water-
shed encompasses about 1300 km2 and the subsurface watershed about 2700 km2.
The average flow is 118 m3/s.
The main structures of the Trebinjica water system are Lake Bilecko (i.e. the
Bileca Reservoir) behind the Grancarevo Dam, and the Gorica Dam and reservoir
downstream. Active operating HPPs are the Trebinje I (180 MW), Trebinje II (8MW),
Capljina (420MW) and Plat (Dubrovnik, 210 MW). Average river flows at the
Grancarevo (Figure 1.11) and Gorica dam sites are 74.3 m3/s and 85.6 m3/s, respec-
tively. Losses from the Bileca Reservoir are negligible, while losses from the Gorica
Reservoir are approximately 57% of the average river flows. These losses appear
mostly downstream from the Gorica Dam and represent a guaranteed ecological flow
to the regulated Trebinjica River (Figure 1.12).
The number of artificial reservoirs in Montenegro is small in comparison with the
hydropower potential. The total capacity of these reservoirs amounts to slightly more
than 1 109 m3. With respect to the total amount of surface water (about 14 109 m3/year)
in the territory of Montenegro, this amounts to about 7% (Hrvacevic, 2004). In the
Adriatic basin, the reservoirs in the Nikicko Polje (Krupac, Slano and Vrtac) have
been formed on the River Zeta, while the Liverovici Dam controls the flow of the
Gracanica River. Water from the reservoirs is utilised by the Perucica HPP (307 MW).
All the reservoirs in the Nikicko Polje have been built in highly karstified rocks. The
Slano and Vrtac required intensive and expensive anti-infiltration works. The grout
curtain along the southern rim of the Slano is one of the longest in the world. It has a
16 Karst without Boundaries
length of 7,011 m, depth of 57 m and surface of 396 122 m2. The current hydropower
capacity of the main Albanian plants is 1750 MW, and most of the dams were built
on the Drini catchment.
The Transboundary Diagnostic Analysis (TDA) was conducted in the period 2011
2013 by the DIKTAS Project Team in accordance with the GEF guidelines. The TDA is
based on a substantial regional analysis that was required in order to fully understand
the context of transboundary issues. The regional analysis was particularly important
given the complexity of the karst environment and regime and interconnectivity of
karst aquifers.
The Project Team was organised into four working groups, reflecting the main
issues of the regional analysis:
)
1 Hydrogeology of the Dinaric karst;
2) Environmental and socio-economic analysis;
3) Legal and institutional framework and policy; and
4) Stakeholder analysis.
Dinaric Karst Aquifer One of the worlds largest transboundary systems 17
Figure 1.12 Channeled Trebinjica riverbed and highly karstified bedding planes, (photo Z. Stevanovic).
One of the main DIKTAS outputs in terms of the regional hydrogeological char-
acterisation is the GIS based digital Hydrogeological map of the Dinaric Karst region
(Figure 1.13). Its creation involved harmonisation of data, classifications, methodolo-
gies, reference systems, projections and semantics. The map was used as a basis for the
development of various thematic maps for the environmental and socio-economical
assessment and represents an important tool for various further analyses and karst
aquifer management.
Figure 1.13 Hydrogeological map of the Dinaric karst (DIKTAS database; http://dinaric.iwlearn.org).
Dinaric Karst Aquifer One of the worlds largest transboundary systems 19
Through the TDA the DIKTAS project was focused on selected and prioritised
transboundary karstic aquifers (TBAs), examining current and potential issues of
concern. Based on five criteria (importance, representativeness, data availability,
issues of concern, relevance), eight TBAs were initially selected for detailed anal-
ysis: Una, Krka, Cetina, Neretva, Trebinjica (all shared by Croatia and Bosnia
Herzegovina), Bilecko Lake, Piva (Bosnia & Herzegovina and Montenegro) and
Cijevna/Cemi (Montenegro and Albania). Six of these TBAs belong to the Adriatic
Sea catchment area while two are part of the Black Sea basin (Una and Piva). The
TBAs comprise a total surface area of 12000 km2, which is approximately 10% of
the entire study area of the four project countries which share Dinaric karst. The
surface area of individual TBAs varies from 668 km2 (Krka) to 3,455 km2 (Cetina).
The later diagnostic analysis focussed on six more problematic TBAs, excluding
Krka and Piva.
A delineation of the aquifer surface area was the first step in the hydrogeological
analysis of each TBA. Further analysis included the characterisation and development
of conceptual models. Groundwater budgeting of TBAs created a base for the assess-
ment of groundwater reserves and availability, as well as for proposals and measures
aiming to ensure sustainable development of TBAs. Some of the selected TBAs, such
as Una, Neretva or Cijevna/Cemi, are of particular importance because they represent
parts of designated protected zones, or wetlands, or the habitat for endangered species
(Figure 1.14).
Three major shortcomings were identified in the region:
The TDA indicated that water extraction was still far below the aquifers replenish-
ment potential, and there is no evidence of significant over-exploitation in the studied
TBAs. For instance, in the case of Cetina and Neretva TBAs the average extraction of
groundwater is ten times less than the total minimal discharge of the springs (dynamic
reserves). However, shortage of water is locally in evidence during summer and early
autumn months which coincides with increased demands during the tourist season.
20 Karst without Boundaries
Figure 1.14 The sablacustrian springs in the Skadar basin (Montenegro): the Karuc, previously planned
for water supply of the coastal area. Skadar (Shkoder) Lake is included in the Ramsar
convention for protected wetlands.
and preventive measures. The Dinaric karst aquifer system areas require specific solu-
tions and compromises in land use planning, and the protection of nature and water
resources. Some of the important springs with catchments shared by the Dinaric coun-
tries are included in the Strategic Action Plan as demonstration sites for the applica-
tion of methodology and design of sanitary protection zones.
The actual monitoring of groundwater in the region and in the studied TBAs is
far from satisfactory. Only in Croatia has the characterisation of groundwater bodies
been completed and monitored in accordance with EU Water Framework Directive.
One of the tasks of the DIKTAS project was to prepare a proposal for the creation
of a new monitoring network which will fully respect karst specific behaviour and
include local water users (waterworks, dams, irrigation, industry). The Cijevna/Cemi
TBA has been identified as the most problematic in terms of available data on water
resources, and installation of a modern surface and groundwater monitoring network
for observation of climate elements has been proposed.
In all four countries there are on-going efforts for transposition of the funda-
mental principles, objectives and measures from the EU Water Framework Directive,
(2000/60/EC) and the Groundwater Directive (2006/118/EC) in national legislations.
Although the polluter pays principle and the principle of recovery of the costs are
promoted in national legislative documents, the principle of cost recovery is not fully
transposed either in national regulations or in water management practices, with
regards to implementation of the environmental and resource costs in water pricing
policies. There is no legal or policy document in any of project countries which ade-
quately defines and prescribes the integration of environmental and resource costs
into development of pricing policies. It should be noted that the main shortcoming
of the legislative framework in all countries is an underdeveloped system of by-laws
and insufficient implementation of present legislation due to lack of human resources
and financial means for fulfilling legal and policy requirements. Due to the lack of
clear development strategies, programmes and plans for water management issues,
the Dinaric Karst region cannot be considered as an example of successful implemen-
tation of these principles.
Based on TDA a Strategic Action Plan which includes different common activities,
proposals for legal and institutional reforms and harmonisation of legislation has been
prepared. In order to support sustainable utilisation of groundwater in the Dinaric
region, the Strategic Action Plan is focusing on several priority actions: improvement
of the quality of water, elimination of sources of pollution, improvement of minimal
discharges, assurance of ecological flows, and establishment of proper water monitor-
ing systems (Table 1.1). Conceptual proposals for investment in all of these areas and
at designated sites have also been prepared as a part of the Plan.
The DIKTAS Strategic Action Plan adheres to two key environmental manage-
ment principles. These are:
The vision guiding the long-term water resources and ecosystem quality objectives is:
To achieve joint sustainable and equitable use and protection of Dinaric karst
aquifer system.
22 Karst without Boundaries
Table 1.1 Proposed actions and suggested TBAs where they can be performed.
Groundwater Quantity
Objective A: Ensure sufficient groundwater availability in dry periods, especially
for water supply and to support environmental flow.
Groundwater Quality
Objective B: Maintain and improve (where required) quality of karst ground
water in the Dinaric region.
Protection of Groundwater Dependent Ecosystems (GDE)
Objective C: Ensure protection of GDE, specific features and their ecosystem
services for the future.
Equitable Use
Objective D: Support equitable use of groundwater resources.
Capacity Building
Objective E: Raise awareness and capacity building related to karst water and
dependant ecosystem.
Dinaric Karst Aquifer One of the worlds largest transboundary systems 23
Regional management actions and desired outcomes for reaching the expected
Water Resources and Ecosystem Quality Objectives (WREQO) are concentrated on
four classes of actions to provide structure to the future implementation:
Regular exchange of data represents one of the key steps toward sustainable
transboundary aquifer management. This fact is recognised also in international law
and acknowledged in the UN Resolution 63/124 the Law of Transboundary Aquifers
(Article 8 Regular exchange of data and information). The DIKTAS countries agree and
commit to exchange information among themselves at the regional or bilateral levels
and prepare an instrument for the legal basis (e.g. Memorandum of Understanding,
or other agreements). DIKTAS countries will exchange collected data through the
national institutions responsible for managing of National Water Information System
and under supervision of the Consultation and Information Exchange Body.1
The Consultation and Information Exchange Body is intended to create conditions
for the development of sustainable and productive consultation mechanisms for infor-
mation exchange and collaboration among the four DIKTAS countries aiming at, among
other, ensure equitable use of groundwater resources in the whole DIKTAS region. The
Consultation and Information Exchange Body will deal with, but not be limited to:
elaboration of the aquifer acts, consideration of improved water supply in critical peri-
ods by proper management and technical solutions; prioritisation of the water end-users
related to water availability and maintaining common information system.
It was concluded that all countries have legislation in place for groundwater pro-
tection which is, to a greater or lesser extent, implemented in each country. However,
due to different criteria for delineation of protection zones and different prescribed
protection measures within these zones, it is not possible to establish and implement
efficient groundwater protection in the areas of transboundary aquifers based on the
existing legislation. The Strategic Action Plan proposes to develop common criteria
based on the existing ones for delineation of protection zones and common protection
measures within these zones.
The establishment of a regional expert working group from experts in the fields
of hydrogeology, water protection, economy and legislation is also envisaged. The
regional expert working groups main tasks should be to determine the criteria for
delineation of protection zones, define programmes for additional surveys, follow up
monitoring network establishment and data exchange, and similar.
Finally, one of the most important objectives is to raise awareness and capac-
ity building related to karst water and dependant ecosystem. The environmental
1 The signing of multilateral agreement between DIKTAS countries should provide a base for establish-
ing the Consultation and Information Exchange Body. Signed Multilateral Agreement on establishment
ofthe Consultation and Information Exchange Body will, among others, represent the acceptance by the
DIKTAS countries to exchange data, which will be included in the Common Information System.
24 Karst without Boundaries
awareness campaign will be based on using the best practices and further elabora-
tion of new public awareness concepts in accordance with the project objectives.
The responsibility of the implementation of the global awareness campaign will be
under the Consultation and Information Exchange Body Secretariat. Promotional
materials in printed, and in electronic form and publications will be an important
channel of communication. Formats will be various: leaflets, newsletters, brochures
and should be distributed at all major events press conferences, school lectures
and meetings.
1.5 CONCLUSIONS
Following Cvijics research, a large number of specialists have further improved the
knowledge of the Dinarides in terms of hydrology, geomorphology, geology and
hydrogeology. Today, more than a hundred years after the initial research, the Dinaric
karst is relatively well investigated on a regional scale, but due to the complexity and
intensity of the karst aquifer regime, detailed survey and systematic monitoring have
to be further improved. The DIKTAS project contributed to better understanding of
complex transboundary inter-linkages and the importance of sustainable and equita-
ble use of water resources.
Due to its historical importance for the development of karst science an initiative
has been taken to include the Dinaric region and its selected areas in the UNESCO list
of World Heritage Sites. This reflets its exemplary karst development with numerous
geo-heritage sites, several national parks and protected areas such as those under
Ramsars Convention, endemic species which inhabit underground world of caves and
abundant groundwater resources. If this proposal comes to reality this will further
strengthen activities on nature and water protection and raise awareness of the local
population of the importance of sustainable use of natural resources.
REFERENCES
Bonacci O. (1987) Karst Hydrology; with special reference to the Dinaric Karst. Springer-
Verlag, Berlin.
Boicevic S. (1966) Caves, potholes and ponors with water in Dinaric karst area (in Croatian).
Kr Jugoslavije, Jug. Akad. Zn. i Um. Zagreb, 6, 105136.
Cvijic J. (1893) Das Karstphnomen. Versuch einer morphologischen Monographie.
Geographischen Abhandlung, Wien, V(3), 218329.
DIKTAS database; http://dinaric.iwlearn.org. Last visited on 10/05/2015.
DIKTAS, 2014: Strategic action plan, Trebinje.
Eftimi R. (2010) Hydrogeological characteristics of Albania. AQUAmundi. 1, 7992.
Ford D. (2005) Jovan Cvijic and the founding of karst geomorphology. In: Stevanovic Z &
Mijatovic B. (eds.): Cvijic and karst, Board on karst and spel. Serb. Acad. of Sci. and Arts,
Belgrade, 305321.
Herak M. (1972) Karst of Yugoslavia. In: Herak, M. and Stringfield, V.T. (eds.), Karst:
Important Karst Regions of the Northern Hemisphere. Amsterdam, Elsevier. 2583.
Hrvacevic S. (2004) Resources of surface water in Montenegro. Elektroprivreda Crne Gore.
Podgorica. 331 p.
Dinaric Karst Aquifer One of the worlds largest transboundary systems 25
Komatina M. (1983) Hydrogeologic features of Dinaric karst. In: Hydrogeology of the Dinaric
Karst. Mijatovic B (ed.). Spec. ed. Geozavod, Belgrade. 4558.
Kreic N. (1988) Karst and caves of Yugoslavia, Naucna knjiga, Belgrade, 149 p.
Kukuric N. (2011) Assessment of internationally shared Karst aquifers: example of Dinaric karst
aquifer system, In: Polk, J. and North, L, Proceedings of the 2011 International Conference
on Karst Hydrogeology and Ecosystems. Environmental Sustainability Publications. Book 2.
Mee S., Aliaj Sh. (2000) Geology of Albania. Gebrder Borntaeger. Berlin-Stuttgart.
Mijatovic B. (1983) Karst poljes in Dinarides. In: Hydrogeology of Dinaric karst. Field trip
to the Dinaric karst, Yugoslavia, May 1528, 1983. (ed. Mijatovic B). Geozavod and
SITRGMJ, Belgrade, p. 6984.
Mijatovic B. (1984) Hydrogeology of the Dinaric Karst. International Association of
Hydrogeologists, Heise, Hannover. Vol. 4.
Milanovic P. (1979) Hydrogeology of karst and methods of investigation (in Serbian), HET,
Trebinje, 302 p.
Milanovic P. (2000) Geological engineering in karst. Zebra Publishing Ltd., Belgrade. 347 p.
Milanovic P. (2005) Water potential in south-eastern Dinarides. In: Stevanovic Z. & Milanovic
P. (eds.): Water Resources and Environmental Problems in Karst CVIJIC 2005, Spec. ed.
FMG. Belgrade, 249257.
Milanovic P. (2006) Karst of eastern Herzegovina and Dubrovnik littoral. ASOS, Belgrade,
362 p.
Milanovic S. (2007) Hydrogeological characteristics of some deep siphonal springs in Serbia
and Montenegro karst. Environ. Geol. 51(5), 755759.
Peka Z., Jolovic B., Radojevic D., Pambuku A., Stevanovic Z., Kukuric N., Zubac Z. (2012)
Unstable regime of Dinaric karst aquifers as a major concern for their sustainable utilization.
Proceedings of 39 IAH Congress, Niagara Falls, (CD publ.).
Radulovic M. (2000) Karst hydrogeology of Montenegro. Sep. issue of Geological Bulletin, vol.
XVIII, Spec. ed. Geol. Survey of Montenegro, Podgorica, 271 p.
Roglic J. (1972) Historical review of morphological concepts. In: Herak, M. and Stringfield,
V.T. (eds.), Karst: Important Karst Regions of the Northern Hemisphere. Amsterdam,
Elsevier Publishing Company. 117.
Stevanovic Z., Mijatovic B. (2005) Cvijic and karst / Cvijic et karst, Monograph: Spec. ed of
Board of Karst and Speleology SANU, Belgrade, 405 p.
Stevanovic Z., Eftimi R. (2010) Karstic sources of water supply for large consumers in
south-eastern Europe sustainability, disputes and advantages, Geologica Croatica, 63(2),
179186.
Stevanovic Z., Kukuric N., Treidel H., Peka Z., Jolovic B., Radojevic D. Pambuku A. (2012)
Characterization of transboundary aquifers in Dinaric karst A base for sustainable water
management at regional and local scale. Proceedings of 39 IAH Congress, Niagara Falls,
(CD publ).
arin A. (1983) Hydrogeologic regional classification of the karst of Yugoslavia. In:
Hydrogeology of the Dinaric Karst. Mijatovic B (ed.). Spec. ed. Geozavod, Belgrade, 3544.
UN Resolution 63(124), The law of transboundary aquifers.
This page intentionally left blank
Chapter 2
ABSTRACT
Karst aquifers may also be transboundary, i.e. crossing political or administrative borders.
Typical karst hydrogeology includes a lack of correspondence between the surface and subsur-
face limits and difficulties in identifying hydraulic boundaries and volumes. A large amount
of high-quality drinkable water comes from karst aquifers, with this volume increasing in the
future. There is an urgent need to understand karst, and safeguard it and its resources, without
being limited by administrative borders. Karst management requires cooperation between the
people living on the land, to transfer and disseminate research outcomes, to create awareness of
the fragile environment, to build resilience and readiness to cope with natural hazards and the
changes they cause, and to minimise disturbance.
Karst is of huge importance for groundwater storage and transport in aquifers which
have been used for many decades for supplying drinking water (Figure 2.1). About
one quarter of the global population is completely or partially dependent on drinking
water from karst aquifers, with karst regions representing about 10% of the Earths
continental area (Ford & Williams, 2007; Hartmann et al., 2014). The percentage of
drinking water from karst is expected to increase in the future. However, the valuable,
high-quality, freshwater resources contained in karst aquifers are extremely vulnera-
ble to contamination.
Karst aquifers are often transboundary, i.e. crossing political or administrative
borders (Figure 2.2). To define the limits of a water catchment, and to understand
whether the likely surface limits correspond to those that are present underground,
is a very difficult task in karst lands, at the same time representing one of its main
peculiarities (Gunn, 2007). A typical element of karst hydrogeology is the lack of
correspondence between the surface and the subsurface limits, which also has conse-
quences in terms of politics, planning and management of territories and of natural
resources.
In the great majority of the natural environments, catchment boundaries are
defined by surface morphology, by outlining ridges and watersheds, dividing a cer-
tain area into several basins. Knowing the amount of rainfall that falls in a specific
28 Karst without Boundaries
Figure 2.1 The Ombla spring, source of drinking water for the town of Dubrovnik, in Croatia.
A) B)
Figure 2.2 The Reka River originates in the Classical Karst. After a surface course on flysch deposits, it
sinks underground at the kocjanske Jame, characterised by a 200 m high, collapsed doline
(A), and follows a course of some 40 km, before coming to the surface in Italy, near Trieste,
at the Timavo resurgences (B), to eventually flow into the Adriatic Sea.
watershed, as well as its size and geological characters, it is possible to estimate the
amount of water collected in the corresponding discharge area.
Water circulation in karst differs from other types of terrain. This is due to limited
surface runoff, balanced by rapid infiltration of water underground through the com-
plex network of discontinuities within the rock mass, and regulated by the role of the
epikarst. Water sinks at a certain site, and, due to complex underground systems made
of conduits and caves, can be transported to another, nearby, watershed. The only way
to be sure of the course of water in karst is to follow it underground, where possible,
or using dye tracers (Goldscheider & Drew, 2007) (Figure 2.3). Locating the karst
catchment boundaries represents a highly complex goal, not least because of the great
variability that may be recorded in time. It has been proved that in conditions of high
How confident are we about the definition of boundaries in karst? 29
Figure 2.3 Dye tracing test within a cave in southern Italy (photo: N. Damiano, Alburni Exploration Team).
groundwater levels, fossil and otherwise inactive karst conduits become activated, thus
originating overflow from one catchment to those nearby (De Waele & Parise, 2013).
For the sustainable development of transboundary karst terrains, the main
actions should address facilitating a fair and sustainable use and management of the
transboundary water resources, and safeguarding the unique groundwater ecosystems
that are present in karst from natural and man-made hazards (Parise & Gunn, 2007;
Gutierrez, 2010; Parise, 2010; De Waele et al., 2011). Such hazards now also include
climate change. These goals should be pursued by taking advantage of the lessons
learned from the ancient communities that used scarce available amounts of water
(Parise & Sammarco, 2015) or from the practices for water use in arid and semi-arid
lands (Stevanovic et al., 2012).
Dagnino Pastore and co-workers (2013) point out several negative tendencies
regarding the management of water resources:
Decision makers and land use planners, attempting to achieve sustainable develop-
ment, must take into account the multiple aspects of water resources management
(van Dijk & Zhang, 2005). The delicate balance existing in karst, in particular with
regard to the transboundary karst water resources, is not easily maintained. There needs
to be specific actions to safeguard the high-value biodiversity of the area (Culver &
Pipan, 2009).
The extreme heterogeneity of karst, with a high number of variables in different
fields, each one of them changing in both time and space, makes analysis complex.
A multi-disciplinary approach is required, and understanding the interaction between
groundwater and surface water, as well as their influence on biota, both at the ground
surface and underground, are crucial points to be addressed.
Several examples can be recalled to outline the delicate and fragile importance
of transboundary karst. In Europe, the lakes Ohrid and Prespa are located in the
Balkan Peninsula (Figure 2.5). Lake Ohrid is shared by Macedonia and Albania, and
Lake Prespa by Macedonia, Albania and Greece (Popovska & Bonacci, 2007). Many
caves are present in the surrounding areas, and the lakes are a well known site for
biodiversity and cultural heritage (Lake Ohrid was declared a World Heritage Site by
UNESCO in 1980). Lake Ohrid is fed by the waters coming from Lake Prespa, that
has no surface outflow (Amataj et al., 2007). Hydrologically, the two lakes should be
managed as a transboundary water resource. However, so far, no coordinated water
resources management has been established.
How confident are we about the definition of boundaries in karst? 31
A) B)
C)
Figure 2.4 Karst springs: A) the Auso, one of the main springs at the foothills of the Alburni Massif
(Southern Italy); B) SyriiKalter, a beautiful spring in southern Albania; C) the main spring
of the Unica River, coming out from the Planinska Jama at the southern edge of Planinsko
Polje, in Slovenia.
N 1
2
3
5
6
Ohrid Lake
MACEDONIA
Macro
Prespa
Lake
ALBANIA
Micro
Prespa GREECE
Lake
0 20
km
Figure 2.5 Sketch of the Ohrid and Prespa Lakes, at the boundaries among three countries (modified
after Popovska & Bonacci, 2007). Key: 1) lake; 2) karst spring; 3) cave; 4) state boundary;
5) water course; 6) limit of the karst catchment (as identified by Popovska & Bonacci, 2007).
groundwaters are the most important source of freshwater. The collection and
analysis of data and preparation of hydrogeological maps on transboundary aquifers
are key objectives (INWEB, 2008).
Mexico is one country that is interested in transboundary water management,
in karst as well in non-karst settings. At its north western borders, the state of Baja
California is located within a large area characterised by arid and semi-arid climate.
The area has become one of the most important economic regions in Mexico despite
water scarcity. About 4 million people live there (the main municipalities are Mexicali,
Tecate, Tijuana, and San Luis Ro Colorado), using the Colorado River as the main
water source by agreement with the American Government in accordance to the 1994
International Waters Treaty.
The situation of the transboundary watershed with the United States is of even
greater concern in the face of the possible effects produced by climate change, espe-
cially because droughts are expected to occur in the future. The problem has even
How confident are we about the definition of boundaries in karst? 33
become more acute due to a significant increase in urban and industrial demand
(Romn Calleros et al., 2010).
On the southern Mexican border (shared with Guatemala and Belize), there is
another transboundary aquifer. Water availability is not a problem in terms of quan-
tity and quality, but there are differences between the individual countries as regards
the basic water and sanitation services. Although the area contains six transboundary
river basins, shared cooperation of water among the countries is virtually absent.
In the Dominican Republic, the karst region of Los Haitises, covering an area of
about 1.6 km2, is an important geomorphologic landscape of limestone domes, with
large volumes of good-quality water. This karst ecosystem is of extreme importance
for water supply in the area, and potentially represents one of the main alternatives
for the future water supply for part of the city of Santo Domingo and the eastern part
of the country. However, the subterranean waters are at high risk of pollution from a
number of planned anthropogenic actions.
About 67% of the Cuban territory is mostly karstified carbonate rocks. Being an
island there is always risk of seawater intrusion (Hernandez et al., 2013). The delicate
balance among the different types of waters can easily be affected by anthropogenic
actions (Molerio Len & Parise, 2009).
The term resilience was initially proposed in the field of ecology as a core concept
within ecosystems (Holling, 1973) and has been widely used in recent years, as the
ability of a system to absorb perturbation or disturbances. Resilience is the capacity
to resist and recover from losses (Kleina et al., 2003). One of the more widespread
definitions of resilience nowadays includes: i) response to disturbance; ii) capacity to
self-organise; and iii) capacity to learn and adapt.
Since the social and environmental consequences of degradation (including pollu-
tion events) are increasing, and cause high costs to society, there is the need to involve
stakeholders and the population in strategies to manage the impacts. This is especially
required in heavily populated areas, or in those sectors in the proximity of big cities
and metropolitan areas, where the natural and anthropogenic factors are mixed (Serre &
Barroca, 2013).
In karst, the concept of disturbance produced by man to the natural environment
has been analysed by means of the Karst Disturbance Index (van Beynen & Townsend,
2005; North et al., 2009). The Karst Distribution Index takes into account a number
of indicators, subdivided into five different categories (geomorphology, atmosphere,
hydrology, biota, cultural), to evaluate the disturbance to the karst environment deriv-
ing by anthropogenic actions. The index has been applied to different settings, from
European karst areas, to selected states or regions in the Americas. As the main out-
come from these applications, it appears that knowledge of the main features of the
karst environments, encompassing many different fields and disciplines of interest,
is fundamental for a proper understanding of the changes occurring, and for linking
such changes to specific actions by man or to variability of other factors (i.e. climate
change).
34 Karst without Boundaries
Among the other indices that have also been designed for use in karst environ-
ments are the Karst Sustainability Index (van Beynen et al., 2012), and the Priority
Management Index (Angulo et al., 2013). The first is a standardised metric of
sustainable development practices in karst settings. It uses predetermined targets
to ascertain the overall sustainability of a karst region, by taking into account
25 indicators related to the fields of environment, economy, and society. The Karst
Sustainability Index eventually provides a measure of sustainability in karst areas.
The second index (by Angulo et al., 2013) is a holistic approach to analyse the
level of disturbance in a protected karst area. Use of two standardised indices
(zonal Karst Significance Index, and zonal Karst Disturbance Index) allows the
most disturbed sectors to be identified. The two indices can be combined into a
Priority Management Index, in order to highlight the main management needs of
the protected area.
Another is the evaluation of disturbance in caves (Harley et al., 2011), where the
Cave Disturbance Index addresses deficient cave management strategies. The index
was based on cave resource data collected during an in situ inventory of 36 terrestrial
caves in Florida. The approach was later followed by Van Aken et al. (2014) using a
GIS-based model to predict disturbance for the same sample of caves. This approach
aims at providing land managers or stakeholders with tools to conduct an in situ
inventory of a cave, in order to better focus both management strategies and conser-
vation efforts.
The main positive effect from all these approaches is that the peculiarity of karst
has been partly acknowledged, together with the need of specific actions for study,
management, and decision, based upon the main features of this highly vulnerable
environment. Recognition of the uniqueness of karst is a crucial point to be reached
in the process of creating an environmental awareness among the communities living
in karst lands.
Repeated episodes of mismanagement in karst, with events of pollution and con-
tamination of karst aquifers, both at the surface and underground, or as post-war
scenarios (Cal & Parise, 2009), highlight how fragile karst is.
The concept of resilience could be applied to floods (Figure 2.6) in karst (Parise,
2003; Gutierrez et al., 2014; Jourde et al., 2014). In general, floods are the most
A) B)
Figure 2.6 Effects of floods in karst environments: A) inundated area in the aftermath of a hurricane
in Cuba in 2009 (photo courtesy of H. Farfn Gonzalez); B) flash flood in the city of Bari
(Apulia, Southern Italy) (after Mossa, 2009).
How confident are we about the definition of boundaries in karst? 35
common type of natural hazards in the world (Liao, 2012). It has been estimated
that they constitute 40% of the natural disasters, and have caused 8.6% of the total
casualties, and 80% of injured people from all types of natural hazards in the last
25 years. In a typical tropical karst landscape, floods are often associated with changes
in land use, through implementation of inadequate agricultural practices, expansion
of urban areas, bad constructive practices, and other human actions inappropriate for
the environment (Farfan Gonzalez et al., 2009).
Floods are also an interesting issue as regards the possible different uses of water.
Especially in regions where there is a high contrast between the wet and dry seasons
(lack of surface water resources, with intervening short periods of abundant water, if
not floods), a proper regulation and distribution of the amount of water available are
needed. Some of the engineering works that promote constant use of water throughout
the year are damming the rivers, and impounding reservoirs. In the case of transbound-
ary rivers, and of rivers shared by more than one country, since dams modify the flow
of rivers, they might create conflicts.
The regulatory capacity of a reservoir could be used to buffer the impact of excep-
tional rises in water level, thus contributing to flood amelioration. Such actions, how-
ever, should be carefully planned and managed, in order not to disturb the ecological
need for flow of rivers.
The same applies to other engineering works, as outlined by the case of the Ombla
spring, in Croatia. The Ombla spring is the source of drinking water for the town of
Dubrovnik (Figure 2.1), with an average discharge of some 24m3/s. The spring is
located at the tectonic contact between karstified Mesozoic carbonates and impervi-
ous flysch sediments, with these latter acting as a hydrogeological barrier (Milanovic
et al., 2014). Construction of a hydroelectrical power plant has been planned, which
raised several doubts in the scientific community and the local populations. The plant
would consist of an underground dam (spanning from 280m below sea level to
135m above sea level) and a power plant.
An interesting example of the different uses of water, even linked to exploitation
activities in karst, is represented by the Pertosa cave, in Campania, Southern Italy.
The cave, about 3 km long, is located at the north east foothills of the Alburni Massifs,
the most important karst area in Southern Italy, hosting several hundred caves
(Del Vecchio et al., 2013). It became a show cave in 1932, only two years after
the opening of the first show cave in the area, the nearby Castelcivita Cave, sit-
uated on the other side of the same carbonate massif. The initial interest in the
cave was archaeological, since at its entrance excavations carried out at the start
of the twentieth century revealed Paleolithic and Neolithic age material (Larocca,
2015). The cave is an active resurgence, hosting a river with discharge in the range
350600 l/s, and a dam for hydroelectric power was built at the site in the 1920s
(Figure 2.7).
2.4 CONCLUSIONS
The peculiarities of karst environments require a specific approach that accounts for
the interrelationships between the surface and the underground environments, as well
as the time and space variability of many of the factors influencing karst ecosystems.
36 Karst without Boundaries
Figure 2.7 Double view of the entrance of the Pertosa Cave and its underground stream: on the left
the cave with the artificial water basin at his highest level held by the dam; on the right with
the open dam and natural water naturally flow (after Larocca, 2015).
Karst aquifers are a valuable natural resource in many transboundary areas, and since
biodiversity in karst represents a remarkable and valuable character, it is necessary to
safeguard this environment, both from the scientific and technical and the administra-
tive and management standpoints.
Many karst aquifers are transboundary. Sustainable development in karst follows
actions that allow growth of the local economy but, at the same time, are respectful of
the peculiarities of this environment. Where mismanagement of karst and degradation
of the natural resources have occurred, efforts should be made towards a resilience
approach, aimed at fully understanding the disturbance caused by man, and at both
recovering the karst landscape and exploiting the capacity to learn and adapt from
past errors or situations.
REFERENCES
Amataj S., Anovski T., Benischke R., Eftimi R., Gourcy L.L., Kola L., Leontiadis I., Micevski E.,
Stamos A., Zoto J. (2007) Tracer methods used to verify the hypothesis of Cvijic about the under-
ground connection between Prespa and Ohrid lake. Environmental Geology 51(5), 749753.
Angulo B., Morales T., Uriarte J.A., Antigedad I. (2013) Implementing a comprehensive
approach for evaluating significance and disturbance in protected karst areas to guide man-
agement strategies. Journal of Environmental Management 130, 386396.
How confident are we about the definition of boundaries in karst? 37
Cal F., Parise M. (2009) Waste management and problems of groundwater pollution in
karst environments in the context of a post-conflict scenario: the case of Mostar (Bosnia
Herzegovina). Habitat International 33(1), 6372.
Culver D.C., Pipan T. (2009) The biology of caves and other subterranean habitats. Oxford
University Press.
Dagnino Pastore J.M., Sturzenegger A., Charreau E.H., Vard O., Bauer C., Bereciarta P.
(2013) Considerations on the situation of water resources in Argentina. In: Jimenez-
Cisneros B., Galizia-Tundisi J. (editors), Diagnosis of waters in the Americas, IANAS Water
Programs, 1789.
Del Vecchio U., Lo Mastro F., Maurano F., Parise M., Santo A. (2013) The Alburni Massif,
the most important karst area of southern Italy: history of cave explorations and recent
developments. In: Filippi M., Bosak P. (editors), Proceedings 16th International Congress of
Speleology, Brno, 2128 July 2013, 1, 4146.
De Waele J., Parise M. (2013) Discussion on the article Coastal and inland karst morphologies
driven by sea level stands: a GIS based method for their evaluation by Canora F., Fidelibus
D., Spilotro G. Earth Surface Processes and Landforms 38(8), 902907.
De Waele J., Gutierrez F., Parise M., Plan L. (2011) Geomorphology and natural hazards in
karst areas: a review. Geomorphology 134(12), 18.
Farfan Gonzalez H., Corvea Porras J.L., Martinez Maquiera Y., Diaz Guanche C., Aldana
Vilas C., de Bustamante I., Parise M. (2009) Impact of the hurricanes Gustav and Ike in the
karst areas of the Viales National Park, Cuba. Geophysical Research Abstracts 11, 1186.
Ford D., Williams P.W. (2007) Karst hydrogeology and geomorphology. Wiley and Sons.
Goldscheider N., Drew D. (editors) (2007) Methods in karst hydrogeology. International
Contributions to Hydrogeology 26, International Association of Hydrogeologists, 264 p.
Gunn J. (2007) Contributory area definition for groundwater source protection and hazard
mitigation in carbonate aquifers. In: Parise M., Gunn J. (editors), Natural and anthropogenic
hazards in karst areas: recognition, analysis and mitigation. Geological Society of London,
Special Publication 279, 97109.
Gutirrez F. (2010) Hazards associated with karst. In: Alcntara I., Goudie A. (editors),
Geomorphological Hazards and Disaster Prevention. Cambridge University Press,
Cambridge, 161175.
Gutierrez F., Parise M., De Waele J., Jourde H. (2014) A review on natural and human-induced
geohazards and impacts in karst. Earth Science Reviews 138, 6188.
Harley G.L., Polk J.S., North L.A., Reeder P.R. (2011) Application of a cave inventory system
to stimulate development of management strategies: The case of west-central Florida, USA.
Journal of Environmental Management 92, 25472557.
Hartmann A., Goldscheider N., Wagener T., Lange J., Weiler M. (2014) Karst water resources
in a changing world: review of hydrological modeling approaches. Reviews of Geophysics
52, doi:10.1002/2013RG000443.
Hernandez R., Ramirez R., Lopez-Portilla M., Gonzalez P., Antiguedad I., Diaz S. (2013)
Seawater intrusion in the coastal aquifer of Guanahacabibes, Pinar del Rio, Cuba. In: Farfan
Gonzalez H., Corvea Porras J.L., de Bustamante Gutierrez I., LaMoreaux J.W. (editors)
Management of water resources in protected areas. Springer, Berlin, 301308.
Holling C.S. (1973) Resilience and stability of ecological systems. Ann. Rev. Ecol. Syst. 4, 123.
INWEB (2008) Inventories of transboundary groundwater aquifers in the Balkans, UNESCO
Chair and Network INWEB, Thessaloniki, Greece http://www.inweb.gr/.
Jourde, H., Lafare, A., Mazzilli, N., Belaud, G., Neppel, L., Doerfliger, N., Cernesson, F. (2014)
Flash flood mitigation as a positive consequence of anthropogenic forcings on the ground
water resource in a karst catchment. Environmental Earth Sciences 71, 573583.
Kleina R.J.T., Nicholls R.J., Thomalla F. (2003) Resilience to natural hazards: How useful is
this concept? Environmental Hazards 5, 3545.
38 Karst without Boundaries
Larocca F. (2015) La Grotta di Pertosa (Salerno) e il suo giacimento archeologico. In: De Nitto L.,
Maurano F., Parise M. (editors), Proceedings XXII National Congress of Speleology Euro
Speleo Forum 2015 Condividere i dati, Memorie dellIstituto Italiano di Speleologia, 29,
651656.
Liao K.H. (2012) A theory on urban resilience to floods A basis for alternative planning prac-
tices. Ecology and Society 17(4), 48.
Milanovic P., Stevanovic Z., Cokorilo Ilic M. (2014) Field trip guide: Trebinje, June 2014.
DIKTAS Project, Faculty of Mining and Geology, Department for Hydrogeology, Belgrade.
Molerio Len L., Parise M. (2009) Managing environmental problems in Cuban karstic aqui-
fers. Environmental Geology 58(2), 275283.
Mossa M. (2007) The floods in Bari: what history should have taught. Journal of Hydraulic
Research 45(5), 579594.
North L.A., van Beynen P.E., Parise M. (2009) Interregional comparison of karst disturbance:
West-central Florida and southeast Italy. Journal of Environmental Management 90(5),
17701781.
Palmer A.N. (2007) Cave Geology. Cave Books, Dayton.
Parise, M. (2003) Flood history in the karst environment of Castellana-Grotte (Apulia, south-
ern Italy). Natural Hazards and Earth System Sciences 3(6), 593604.
Parise M. (2010) Hazards in karst. In: Bonacci O. (editor), Proc. Int. Sc. Conf. Sustainability
of the Karst Environment, Dinaric Karst and Other Karst Regions, Plitvice Lakes, Croatia,
2326 September 2009, IHP-UNESCO, Series on Groundwater, 2, 155162.
Parise M., Gunn J. (2007) Natural and anthropogenic hazards in karst areas: an introduc-
tion. In: Parise M. & Gunn J. (editors) Natural and anthropogenic hazards in karst a reas:
Recognition, Analysis, and Mitigation. Geological Society, London, Special Publication
279, 13.
Parise M. & Sammarco M. (2015) The historical use of water resources in karst. Environmental
Earth Sciences 74, 143152.
Popovska C., Bonacci O. (2007) Basic data on the hydrology of Lakes Ohrid and Prespa.
Hydrological Processes 21, 658664.
Romn Calleros J., Cortez Lara A., Soto Ortiz R., Escoboza Garca F., Viramontes Olivas O.
(2010) El aguaen el noroeste. In: Jimnez B., Torregrosa M.L., Aboites A. (editors), El Agua
en Mxico: Cauces y encauces. Academia Mexicana de Ciencias, 479504.
Serre D., Barroca D. (2013) Natural hazard resilient cities. Natural Hazards and Earth System
Sciences 13, 26752678.
Stevanovic Z., Balint Z., Gadain H., Trivic B., Marobhe I., Milanovic S. et al. (2012)
Hydrogeological survey and assessment of selected areas in Somaliland and Puntland.
Technical report no. W-20, FAO-SWALIM (GCP/SOM/049/EC) Project, (http://www.
faoswalim.org/water_reports) Nairobi.
UNDP (2007) Informe de desarrollo humano, 2006. Msall de la escasez: poder, pobreza y la
crisismundial del agua. United Nations Development Programme.
Van Aken M., Harley G.L., Dickens J.F., Polk J.S., North L.A. (2014) A GIS-based modeling
approach to predicting cave disturbance in karst landscapes: a case study from west-central
Florida. Physical Geography 35(2), 123133.
van Beynen P.E., Townsend K. (2005) A disturbance index for karst environments. Environmental
Management 36, 101116.
van Beynen P.E., Brinkmann R., van Beynen K. (2012) A sustainability index for karst environ-
ments. Journal of Cave and Karst Studies 74(2), 221234.
van Dijk M.P., Zhang M. (2005) Sustainability indices as a tool for urban managers, evidence
from four medium-sized Chinese cities. Environmental Impact Assessment Review 25,
667688.
Chapter 3
ABSTRACT
The WOKAM project supported by the International Association of Hydrogeologists and
UNESCO was established in 2012 in order to obtain a better global overview of karst aquifers,
to create a basis for sustainable international and global water resources management, and
to increase the awareness and knowledge concerning these special resources, both among the
public and the decision-makers. The team of international experts is responsible for data col-
lection and evaluation, definition of methodology and creation of the new karst aquifers map
and the associated database. The map, to be completed in 2016, will not only show carbonate
rock and evaporite outcrops, but also display deep and confined karst aquifers, large karst
springs, including thermal and mineral springs, drinking water abstraction sites and selected
caves. The World Karst Aquifer Map is based on the highly detailed Global Lithological Map
(GLiM) and is intended to supplement the existing map of Groundwater Resources of the
World (WHYMAP). The paper presents some examples and large differences between karstic
aquifers of the Alpine system of South East Europe, the Near and Middle East, which have
some of the most karstified and richest water reserves in the world, and the less productive and
developed East African karst.
3.1 INTRODUCTION
According to an often-cited estimation by Ford & Williams (2007) surface and sub-
surface outcrops of potentially soluble karstic rocks occupy around 20% of the plan-
ets ice-free land, but probably not more than 1015% is extensively karstified. The
same authors found that probably more than 90% of the evaporitic rocks anhydrite
and gypsum do not crop out, while this percentage in the case of remarkably soluble
salty rocks is almost 99%. About 25% of the global population is at least partly
supplied by freshwater from these karst aquifers. Owing to their specific hydrogeo-
logical properties, karst aquifers are particularly vulnerable to contamination and dif-
ficult to manage. Furthermore, many karst aquifers are hydraulically connected over
40 Karst without Boundaries
wide areas and thus require transboundary exploration, protection and management
concepts.
For almost half a century, since the Karst Commission was established under the
umbrella of the International Association of Hydrogeologists, through the publication
of books, and organisation of conferences, field trips, seminars, and workshops, it has
made many important contributions to the knowledge of karstic aquifers and ground-
water. One of its most recent collaborative efforts is the implementation of the project
WOKAM, the ultimate goal of which is to create a world map and database of karst
aquifers, as a further development and refinement of earlier maps.
Numerous international karst experts contribute to this project by collecting data
and working on an international Scientific Advisory Board that meets on a regular
basis1. The first results of this project were presented by Scientific Advisory Board
members at several global conferences, including Karst without boundaries held
in Trebinje in June, 2014. There were opportunities to discuss the draft legend, the
design and preliminary content of the associated database, as well as the first sketch
maps for Europe, Asia and North America.
Although karst phenomena were listed and presented in numerous books and encyclo-
pedias (e.g. Gunn, 2004; Culver & White, 2005), and the most attractive and largest
were included on many websites, several attempts were also made to present various
karst features on the map and in the common database. These activities took place
largely in the USA (Epstein et al., 2001; Hollingsworth E. et al., 2008; Veni, 2002).
The sketch maps showing regional distribution of carbonate rocks are also presented
in Ford and Williams book Karst Hydrogeology and Geomorphology (2007). The
same map with a somewhat better resolution is available at the web address of the
University of Auckland, New Zealand http://www.sges.auckland.ac.nz/sges_research/
karst.shtm.
After completion of WHYMAP, the map showing the groundwater resources of
the World (Richts et al., 2011), the Past-President of IAH, Willi Struckmeier, proposed
during the karst conference held in Malaga in 2010 that a karst map should be pre-
pared to complement the existing global groundwater resources map. The concept
was accepted and an initiative team under the coordination of Nico Goldscheider
from Karlsruhe Institute of Technology was formed and started to work. From the
beginning it was agreed that one main output would be a GIS-based map in scale
1:10 Million, while final hard (paper) copy will be at a scale of 1:25 Million, with
several layers and a connected database.
1 WOKAM Scientific Advisory Board consists of the following experts (in alphabetical order): Augusto
Auler (Brazil), Michel Bakalowicz (France), David Drew (Ireland), Nico Goldscheider (Germany, Project
leader), Guanghui Jiang (PR China), Jens Hartmann & Nils Moosdorf (Germany, GLiM Team), Andrea
Richts (Germany, WHYMAP Team), Zoran Stevanovic (Serbia) and George Veni (USA). Zhao Chen, PhD
student at KIT, is compiling the map, with support of a student assistant, Franziska Griger. Alexander
Klimchouk, Art Palmer, Paul Williams and Daoxian Yuan contributed by exchange of ideas during the
project preparation.
WOKAM The world karst aquifer mapping project 41
To make such a thematic map on a very small scale, appropriate answers to very
specific issues are required. Some of them are:
1. The mapping of rock distribution is based on the highly detailed Global Lithological
Map (GLiM), which was assembled from 92 regional lithological maps of the high-
est available resolution (Hartmann & Moosdorf, 2012).
2. Deep and confined karst aquifers are identified on the basis of regional literature
and geological sections and their approximate boundaries are shown on the map.
3. WOKAM GIS database includes large karst springs, thermal springs, submarine
springs, major drinking water abstraction sites and important caves, which are all
shown on the map as point symbols.
Figure 3.1 Lithology units in Europe and adjacent areas as a base for the draft WOKAM map.
3.2.3.1 Springs
The numbers in brackets indicate candidates to be included on the map in the database,
while the presentation of selected springs on the final map will be much less detailed:
Permanent karst springs with a minimum discharge > 200 l/s (for the map:
500 l/s2)
2 Later, the Scientific Advisory Board proposed a norm of 2,000 l/s, but still in a flexible manner, so that if
a spring is not reaching the criterion, but it is unique and represents a regionally important karst aquifer
it could be shown on the map. In this case experts opinion is relevant.
WOKAM The world karst aquifer mapping project 43
Temporary or highly variable springs with a maximum discharge > 10 m/s (50 m/s)
Very important submarine springs (experts opinion)
Thermal water resources/systems with total discharge > 100 l/s that are >4C
warmer than the average air temperature (200 l/s and 10C warmer than air
temperature)
Karst springs > 100 l/s with peculiar gas composition, such as CO2 or H2S (200 l/s)
Important wells, well fields, drainage galleries or other water abstraction struc-
tures in karst with a discharge/pumping rate > 200 l/s (500 l/s).
The estimated annual minimum discharge (low flow, baseflow) is thus a main
criterion. It can be determined in three different ways: 1) One or several discharge
measurements or estimations during low-flow conditions; 2) Estimated on the basis
of minimum and mean discharge; 3) Calculation on the basis of long-term discharge
data. But it was also agreed that because of the heterogeneity of the database and the
heterogeneous global distribution of large springs, the legend and selection criteria
could be slightly modified. Accordingly, not all the large springs have to be shown
on the map in terms of their very high density (e.g. Dinaric karst), while some of the
smaller springs that are regionally very important, such as the Jordan spring, although
fulfilling none of criteria, can be shown on the map. Although it is not logical to
exclude some larger springs because there are a lot of them the printing scale is such
that overlapping of objects would be unavoidable. In contrast, some small springs can
be included as an extra item in the legend because of their importance.
3.2.3.2 Caves
The Scientific Advisory Board has established basic ranking criteria by appointing the
following values (total maximum of 10 points): up to 5 points for dimensions (3 for
length and 2 for depth) and up to 5 points for significance (2 for hydrological signifi-
cance, 1 for human use and ecosystems, 2 for regional significance). In principle, caves
with less than 4 or 5 points would not be mapped, and will remain just a part of the
database. The final goal is to have a short list of the most relevant caves for the map.
These rankings also confirmed that caves that are significant in terms of archeology,
cave paintings, fossil remnants, and tourism are less likely to be displayed on the map
than, for instance, a short cave with high hydrologic and regional significance.
Figure 3.2 Study region from Central and South East Europe to Indian sub-continent and eastern
Africa.The lines show countries with information on water points and caves in the project
stage I.
of the scale of the WOKAM map. However, until the end of 2014, involved local
experts provided a significant amount of information and prepared inquiry lists for
springs and caves completed by the Scientific Advisory Board. It might be concluded
that the work has been done successfully, while in a live organism such as the GIS
database, additionally collected data will improve the overall quality of the study and
enable further creation of regional maps at larger scales.
The karstic systems of the region might be divided into two major groups: one
represents aquifers formed in large geosyncline structures (Tethys), recently trans-
formed into a highly deformed orogenic belt consisting of high mountains intersected
by large depressions, and the second, the platform type of karst, slightly deformed and
karstified (Stevanovic, 2015).
The first, the Alpine geosyncline karstic system, has several branches: Carpathians,
Balkanides, Dinarides, Hellenides, Taurides, Zagros, and Himalayan mountain chains
with adjacent areas. It is an area of highly developed karst with some classical karst
regions such as Dinaric karst. Creation of karst is associated with the transgression
of the Tethys Ocean and later intensive orogenesis (intensive uplifting and folding)
repeated in several stages: the two most intensive orogenic phases took place at the
end of the Upper Cretaceous and during the Paleocene (Laramian tectonic phase),
and in Eocene/Oligocene (Helvetian phase). The karst in the Northern Hemisphere
WOKAM The world karst aquifer mapping project 45
3.2.4.1 Springs
The largest number of springs occur in the Dinaric karst (Mijatovic, 1984; Jane etal.,
1997; Kranjc, 1997; Bakalowicz et al., 2003; Stevanovic, 2010; DIKTAS database
http://dinaric.iwlearn.org). Although a more restrictive approach has been applied in
the selection of springs, 53, or 40% of all major springs belong to that aquifer system
across seven countries which nowadays share the water resources. For instance, the
four countries of former Yugoslavia have the highest density of springs: distribution
of large karstic springs which achieve the discharge criterion against territory covered
by karst in Croatia, Slovenia, Montenegro and Bosnia & Herzegovina is equal to or
larger than 1 spring/2000 km2. Probably additional restriction or adaptation of the
legend (grouping of objects) will be required once a full version of the map with water
points/caves is displayed (Figure 3.3).
A considerable number of springs are also characteristic of other countries which
contribute in part to the Dinaric system: In the Albanian part of the Dinarides (Albanian
Alps) there are 2 large springs (Eftimi, 2010), in the Dinarides of western Serbia, 4, and
in FRY of Macedonia, 4. In the southern extensions of Alpides: Pindes and Hellenides,
4 Albanian and 2 Greek springs are included in the WOKAM database.
46 Karst without Boundaries
Figure 3.3 The segment of the digital Hydrogeological map of Dinaric karst prepared under the DIKTAS
project (http://dinaric.iwlearn.org). The boundary area between Bosnia & Herzegovina and
Croatia (between Zrmanja and Neretva rivers) is very rich in karstic springs discharging a
minimum of over 100 l/s (smaller dots) and over 1000 l/s (larger dots), not all of which can
be shown on the WOKAM map.
The Carpathian karst is also rich in groundwater reserves and large springs,
but to a much lesser extent than the southern Alpine branches. The number of large
springs or well-fields included in the WOKAM database is: 1 in Czech Republic, 1 in
Slovakia, 4 in Poland, 3 in Romania (Oraseanu & Iurkiewicz, 2010) and 3 in Serbia
(Stevanovic, 1994). In the Balkanides of Bulgaria there are four large springs pro-
posed for the WOKAM database (Benderev et al., 2005). In the Hungarian part of the
Pannonian basin and its margin there are several fresh water and thermal springs, four
of them are included in the WOKAM list.
The Taurus Mountains in central and southern Turkey are the most karstified
region in the Near East (Gnay, 2010). Most of the 9 selected springs, which
issue from Mesozoic and rarely from Tertiary sediments, belong to that unit. The
six large springs selected are in Israel; among them is the famous Jordan River
source.
In Iraq, five springs met the criteria for inclusion in the WOKAM database. They
belong to the two major geo-structural units, the Taurides and Zagros Mountains,
both extending over the northernmost part of the country along the borders of Turkey
and Iran. They drain the two main karstic aquifer systems, Bekhme of Cretaceous,
and Pila Spi of Tertiary age (Stevanovic & Iurkiewicz, 2004). Similarly, nine major
springs of Iranian karst were selected for inclusion on the WOKAM list. They also
drain limestones of Cretaceous (Sarvak Formation) and Oligocene-Miocene age
(Asmari Formation; Raeisi & Stevanovic, 2010). The evaporitic rocks are connected
to younger formations (Mio-Pliocene, Quaternary).
In India, out of the twelve selected, nine springs are located in Kashmir. The other
three are on the Meghalaya Shillong Plateau.
WOKAM The world karst aquifer mapping project 47
North eastern and eastern African and Arabian plateaus are less karstified and
almost none of these karstic occurrences satisfied the established criterion of min-
imum discharge (Figure 3.4) for the database. However, 3 major spring groups
(oases) have been selected in Egypt, along with one well-field in Somalia and one
well-field in Ethiopia, as representatives of local karst aquifers. Although Eocene
carbonate and evaporitic rocks have major extensions in Somalia and eastern
Ethiopia, the two selected well fields are linked to Jurassic limestones (Stevanovic
etal., 2012).
Concerning the spring discharges, of the 124 springs in this wide region, 28 or
1/5 have a minimal discharge larger than 2000 l/s which represents key evidence of
water availability. The largest is the famous Dumanli spring in Turkey (38 000 l/s),
Figure 3.4 The segment of digital Hydrogeological map of Sanaag karstic area between Somaliland and
Puntland provinces (northern Somalia) prepared under project SWALIM, IV Phase (http://
www.faoswalim.org/water_reports). Karstic springs or well fields situated in Eocene karstic
rocks (all formations on map labeled by E) are relatively dense, but springs that discharge
more than 10 l/s (big dots on the map) are very rare.
48 Karst without Boundaries
nowadays impounded by the Oymapinar reservoir. The second ranked is the Bistrica
group of springs in Albania (including the well-known Blue eye spring) with a mini-
mum discharge of 12 000 l/s. The largest number of springs discharging regularly over
2 000 l/s are in Bosnia & Herzegovina (8) and Turkey (8) followed by Montenegro
(5). Milanovic (2006) noted that springs along the Neretva Valley and in eastern
Herzegovina (Buna, Bunica) are the largest in the Adriatic basin.
3.2.4.2 Caves
The distribution of caves is not always the same as that for springs. Some karstic
countries possess larger speleological objects than water objects. For instance, the
Carpathian karst is specific in its development of large underground cavities formed
along preferred tectonic pathways. As such, 18 caves, or 1/3 of the total proposed for
the WOKAM list, are from that geo-structural unit: Romania is a leader in this group
with 9 caves, followed by Poland (Tatra Mountain) with 3, Czechia (Moravsky kras)
and Hungary with 2, and Slovakia and Serbia with one each.
Slovenia has the most caves in the southern Alpine group with 9 caves proposed
for the WOKAM list. Among them is the famous touristic cave Postojna. Some other
well known caves, or worlds largest potholes, are also selected to be displayed on
the map: the Humpleu-Poienita cave system (Romania, 40 km of explored channels
in length), Demnovsk jaskyne system (Slovakia, 36 km), Amaterska cave (Czechia,
35 km), Wielka Sniena Cave (Poland, 23 km), Cehi 2 pothole (Slovenia, 1500 m in
depth), and Slovacka jama (Croatia, 1300 m deep).
The largest African cave, Sof Omar Holluca in Ethiopia, with 15 km of explored
channels, is included in the list. The largest Indian cave among five proposed for the
WOKAM list (all in Meghalaya) is Krem Liat Prah with 30 km of explored channels.
Of 52 WOKAM caves from the studied region, 11 have more than 20 km of
explored channels, while 6 potholes are deeper than 1000m.
The WOKAM project will be completed in 2016 and the map is scheduled to be
printed in late 2016. Because this chapter was written before the final evaluation of
the submitted data and the Scientific Advisory Boards review, all information is of a
preliminary nature. Figure 3.5 shows a part of the draft map with selected springs of
the south eastern Europe, eastern Mediterranean basin and adjacent areas.
3.4 CONCLUSIONS
The WOKAM project will be completed in 2016 and will result in the World Karst
Aquifer Map, one of the layers of, and supplement to, the existing map of Groundwater
Resources of the World (WHYMAP). It will provide a better global overview of
karst aquifers, and is expected to become a useful tool for sustainable global water
resources management, and to increase public awareness of the importance of these
resources, currently utilised for drinking purposes by some 25% of the worlds population.
WOKAM The world karst aquifer mapping project 49
Legend
spring
non-exposed /
exposed karst aquifer
continuous carbonate
sedimentary rocks
discontinuous carbonate
sedimentary rocks
km
0 250 500 1,000
Figure 3.5 A part of preliminary WOKAM map showing selected springs of the South East Europe
and adjacent areas.
The team of international experts, members of the Scientific Advisory Board work-
ing on the project since 2012, has defined the methodology for data collection and
evaluation, and for the creation of the map and its legend as well as the associated data-
base. As the most feasible solution to present the lithology and corresponding aquifer
systems, the existing Global Lithological Map was chosen. Selection criteria for water
points (springs, thermal springs, well fields, extraction intakes) and caves as single geo-
morphologic karstic forms to be included in the map were defined in accordance with
the preferred scale of 1: 10 Million. However, considering the large differences between
highly karstified areas and rich karstic aquifers, and poorly developed karst with small
water quantity and availability, only two approaches to the selection of objects for the
map and database were found to be feasible: the more restrictive approach in the former
case and the more flexible in the latter. This concept is demonstrated by the examples
from South East Europe, and the Near and Middle East where karstic aquifers belong to
the Alpine system and are the richest in water reserves in the world, and in contrast with
examples from less productive and developed Eastern African karst. In eastern Africa
karstic formations have a large area, but the yield of springs or drilled wells is relatively
low and rarely exceeds 10 l/s. Thus, although they are the main source for local water
supply, they do not meet the WOKAM criteria for display on the map.
ACKNOWLEDGEMENTS
and Middle East: Romeo Eftimi (Albania), Boban Jolovic (Bosnia & Herzegovina),
Aleksey Benderev (Bulgaria), Judit Mdl-Szony (Hungary), elimir Peka (Croatia),
Jerome Perrin (BRGM-France, data for India), Ezzat Raeisi (Iran), Milan Radulovic
(Montenegro), Jacek Rzkowski (Poland), Iancu Oraseanu (Romania), Peter Malik
(Slovakia), Nataa Ravbar (Slovenia) and Gltekin Gnay (Turkey) to collect and pro-
vide valuable data and information. Without the help of all those involved, this work
would not have been complete and productive. We also thank IAH and UNESCO for
provided financial support.
REFERENCES
Bakalowicz M., Fleury P., Drfliger N., Seidel J.L. (2003) Coastal karst aquifers in Mediterranean
regions. A valuable ground water resource in complex aquifers. In: Instituto Geologico y
Minero de Espana Publ., Technologia de la Intrusion de Agua de mar en Acuiferos Costeros:
paises mediterraneos (TIAC). Hydrogeologia y aguas subterraneas, no. 8, Alicante, 2,
p. 125140.
Bonacci O. (1987) Karst Hydrology; with special reference to the Dinaric Karst. Springer-
Verlag, Berlin.
Benderev A., Spasov V., Shanov S., Mihaylova B. (2005) Hydrogeological karst features of
the Western Balkan (Bulgaria) and the anthropological impact. In: Water Resources and
Environmental Problems in Karst CVIJIC 2005, eds. Z. Stevanovic and P. Milanovic. Spec.
ed. FMG. Belgrade, pp. 3742.
Culver D.C., White W.B. (2005) Encyclopedia of Caves. Burlington, MA, Elsevier Academic
Press, 654 p.
Eftimi R. (2010) Hydrogeological characteristics of Albania. AQUAmundi. 1, 7992.
Ford D.C., Williams P.W. (2007) Karst Hydrology and Geomorphology. London, Wiley
Chichester, 2nd ed., 576 p.
Epstein J.B., Orndorff, R.C.,Weary, D.J. (2001) U.S. Geological Survey National Karst map
[abs.]: National Speleological Society Convention Program Guide, 87 p.
Goldscheider N., Chen Zh., WOKAM Team (2014) The world karst aquifer mapping project
WOKAM. [abs.] In: Kukuric N, Stevanovic Z, Kreic N (eds.) Proceedings of the DIKTAS
Conference: Karst without boundaries, Trebinje, June 1115 2014, p. 391.
Gnay G. (2010) Geological and hydrogeological properties of Turkish karst and major karstic
springs. In: Kresic N., Stevanovic Z. (eds.) Groundwater hydrology of springs: Engineering,
theory, management and sustainability, Elsevier Inc., BH, Burlington-Oxford, 479497.
Gunn J. (2004) Encyclopedia of Caves and Karst Science. New York, Fitzroy Dearborn, 902 p.
Hartmann J., Moosdorf N. (2012) The new global lithological map database GLiM: A repre-
sentation of rock properties at the Earth surface. Geochemistry, Geophysics, Geosystems 13
(12): 137. doi:10.1029/2012GC004370.
Hollingsworth E., Van Brahana, Inlander E., Slay M. (2008) Karst Regions of the World
(KROW) (2008): Global karst datasets and maps to advance the protection of karst species
and habitats worldwide. http://pubs.usgs.gov/sir/2008/5023/pdf/06hollings.pdf
Jane J., Car J., Habic P., Podobnik R. (1997) Vodno bogastvo visokega krasa. Geologija d.o.o.,
Idrija, 167 p.
Kranjc A. (1997) Karst hydrogeological investigations in south-western Slovenia. Acta
Carsologica 26(1), 388 p.
Mijatovic B. (1984) Hydrogeology of the Dinaric Karst. International Association of
Hydrogeologists, Heise, Hannover. Vol. 4.
WOKAM The world karst aquifer mapping project 51
Milanovic P. (2006) Karst of eastern Herzegovina and Dubrovnik littoral. ASOS, Belgrade,
362 p.
Oraseanu I., Iurkiewicz A. (2010) (eds) Karst hydrogeology of Romania. Belvedere Publ.
Oradea.
Radulovic, M. (2000) Karst hydrogeology of Montenegro. Sep. issue of Geological Bulletin,
vol. XVIII, Spec. ed. Geol. Survey of Montenegro, Podgorica, 271 p.
Raeisi E., Stevanovic Z. (2010) Springs of Zagros mountain range (Iran and Iraq). In: Kresic N.,
Stevanovic Z. (eds.) Groundwater hydrology of springs: Engineering, theory, management
and sustainability, Elsevier Inc., BH, Burlington-Oxford, 498515.
Richts A., Struckmeier W.F., Zaepke M. (2011) WHYMAP and the Groundwater Resources of the
World 1:25,000,000. In: Jones, J.A.A. (Ed.) Sustaining Groundwater Resources, pp. 159173.
Stevanovic Z. (1994) Karst ground waters of Carpatho Balkanides in Eastern Serbia. In:
Stevanovic Z, Filipovic B. (eds) Ground waters in carbonate rocks of the Carpathian
Balkan mountain range. Spec. ed. of CBGA, Allston, Jersey, pp. 203237.
Stevanovic Z., Iurkiewicz A. (2004) Hydrogeology of Northern Iraq, Vol. 2. General hydroge-
ology and aquifer systems, Spec. Edition TCES, FAO/UN, Rome, 175 p.
Stevanovic Z. (2010) Major springs of southeastern Europe and their utilization, In: Kresic N.,
Stevanovic Z. (eds.) Groundwater hydrology of springs: Engineering, theory, management
and sustainability, Elsevier Inc., BH, Burlington-Oxford, 389410.
Stevanovic Z, Balint Z, Gadain H, Trivic B, Marobhe I, Milanovic S et al. (2012). Hydrogeo
logical survey and assessment of selected areas in Somaliland and Puntland. Technical report
no. W-20, FAO-SWALIM (GCP/SOM/049/EC) Project. (http://www.faoswalim.org/water_
reports) Nairobi.
Stevanovic Z. (2015) Characterization of Karst Aquifer. In: Stevanovic Z. (ed) Karst Aquifers
Characterization and Engineering, Series: Professional Practice in Earth Science. Springer
International Publishing Switzerland, 47126.
Veni, G., April 2002, Revising the karst map of the United States. Journal of Cave and Karst
Studies, 64(1), 4550.
ABSTRACT
This chapter analyses groundwater flow in the Orontes River basin and changes which occurred
in the past forty years as a result of the massive expansion of irrigated lands using the ground-
water resources. The region contains significant karstic water resources supplying springs in
the upper and middle reaches of the basin. Although variations in annual flows are difficult
to assess precisely, there has been a significant decrease since the 1960s. The SyrianLebanese
agreement on sharing of the Orontes water, signed in 1992, focused almost exclusively on
surface water resources. The drilling of wells in Lebanon was restricted only near the Orontes
River. Amendments to the agreement, in 1997 and 2002, restricted the drilling of wells in the
Lebanese section of the basin as a whole. However no restrictions were imposed on ground-
water withdrawals in Syria. Because of the continuity of the main Jurassic-Cretaceous aquifer
between the two countries the latter withdrawals are likely affect groundwater resources in
Lebanon.
4.1 INTRODUCTION
The extensive development of irrigation in the Orontes River basin, since the 1970s,
has substantially disorganised subsurface flows feeding the water resources of this
region, particularly within the area of the Lebanese-Syrian border.
A literature review located the approximate recharge areas of the main springs
and showed the impact of the recent intensive exploitation of the Jurassic-Cretaceous
aquifer on subsurface water flows.
From the upstream Lebanese section of the watershed, surface water carried by
the Orontes river flows toward Syria, as well as significant amounts of groundwater
from the Jurassic-Cretaceous aquifer complex extending at depth into to the Syrian
territory.
This chapter analyses the effects of irrigation development on the dynamics of the
groundwater resources and how this issue was taken into account in the 1994 water
agreement and the 1997 and 2002 amendments. The study was conducted as part of
54 Karst without Boundaries
a research programme supported by the Global Program Water Initiatives of the Swiss
Agency for Development and Cooperation.
The Orontes basin contains significant karstic water resources. These are largely fed
the Orontes River before the extensive development of the irrigation in the past four
decades. In the 1960s, the downstream discharge rate reached almost 100 m3/s at the
Syrian Turkish border, but in the 2000s, it fell to less than 15 m3/s.
The large aquifers supplying the main springs located in the upstream reach of
the basin are thick limestone formations of Jurassic and Cretaceous age (Figure4.1).
Although these formations are not equally fractured and karstified, they contain
groundwater flowing through them and from one to the other via faulting or fractur-
ing, even if they are separated by lower Cretaceous impermeable strata. The scale of
the basin allows the Jurassic and Cretaceous formations to be considered as a unique
complex aquifer, a very large reservoir in hydraulic continuity (Figure 4.2).
In the southern and central parts of the basin, this large karstic aquifer supplies
many springs (Figure 4.3). The annual flow of the main springs is of several m3/s.
Their regime is more or less stable throughout the year because of the very large
water reserves, the high hydraulic conductivity, the well-developed internal drainage
and the extended confinement of this complex aquifer. Recharge to the Jurassic and
Cretaceous aquifer is particularly important in the highest areas of the basin, especif-
ically on Mounts Lebanon and Anti-Lebanon. Where limestone formations outcrop,
the recharge can reach up to 60% of the total precipitation, varying between 750
and 2000 mm per year on the mountain range (Droubi, 2012). Recharge takes place
mainly during the winter season, and lasts until spring due to snow-melt water.
Legend
Spot elevation
Cities
Major springs
Minor springs
Rivers
37N 37N
International
boundary
Major faults
Faults
Watershed
boundary
Hydrogeology
Neogene-
Quaternary
Basalts
Eocene Miocene
Upper Cretaceous
Paleogene
Jurassic Middle
Cretaceous
Ignous rocks
36N C1 Hydrogeologic profile 36N
35N 35N
34N 34N
N
0 12.5 25 50
KM
01.09.2015 / FZ, RG
36E 37E 38E
Remote groundwater discharge from the Jurassic and Cretaceous aquifer has
been much more affected downstream, in particularly in the East of the Al Ghab
plain. Several springs have dried up, as shown in Table 4.1, which provides references
of the sources (Figure 4.1).
Weakly aquifer
Primary sandstone, marls,
greenstone formation
W COASTAL JABAL AZ E
MOUNTAIN GHAB
ZAWIYAH
1250 1250
1000 Naur Nab Al 1000
Shathah Huwayz
500 Orontes 500
C3 0 0
500 500
1000 1000
COASTAL
NW MOUNTAIN SE
BEEQA
NW MOUNT-LEBANON ANTI-LEBANON SE
2000 2000
Hermel s
1500 e 1500
nt
ro
O Ain ez Zarqa
1000 1000
C1
500 500
Figure 4.3 Hydrogeological cross-sections of the Orontes River basin (legend see Figure 4.1).
Groundwater flow in the Orontes River basin 57
Table 4.1 Estimated flows of the main springs or group of springs, in the 1960s and in the late 1990s,
south and central parts of the basin.
1960 2000
Flowlines Groundwater drawdown values
in unconfined aquifer 20 m
in confined aquifer 50 m
Figures 4.4 and 4.5 Schematic diagram of the organisation of the subsurface flow, in the 1960s (left)
and 2000s (right).
50 MCM for MayAug period one) allocated to Lebanon if the river flow exceeds
400 MCM/year at the Hermel Bridge gauging station and 20% of the annual flow if
the discharge volume is less than 400 MCM/year (Sarraf, 2013).
Comair (2013) specified thatthe 1994 Agreement addressed the issue of ground-
water withdrawals with a provision of authorizing pumping from wells drilled before
the signature of 1994 agreement, but prohibiting new wells. The wells allowed were
the ones located within a radius of 1500 m from the center of the source and 500 m
of either banks of the river.
In the absence of dams or diversion works (not mentioned in the agreement), a large
part of the water allocated to Lebanon, in particularly during winter season, cannot
be used. Moreover, because of groundwater withdrawal restrictions, the Lebanese
border village of Qaa and Hermel could not undertake any irrigation development.
Sarraf (2013), at only 21 MCM/year of which 23% is for domestic purposes and the
rest for irrigation (irrigated areas in the Lebanese part of the Orontes basin are offi-
cially 1.703 ha in area).
As the fixed amount of water allocated to Lebanon is of 80 MCM/year, Lebanese
withdrawals are much less than the volume attributed and the situation so far does
not require any specific discussion between the two countries regarding the shared
water part as established in the Agreement. However, sharing the water according the
Agreement is quite complex, involving a calculate for each month, in dry years, of the
water deficit regarding the monthly flow of an average year which will be taken into
account in the following months for reduced withdrawal.
4.6 CONCLUSIONS
This chapter provides an update on recent issues from the hydrogeology of the
Orontes basin and seeks to show how the existing large groundwater resources have
been taken into account in the Lebanese-Syrian Agreement.
Groundwater flow in the Orontes River basin 61
Unlike river flow rate, sharing of groundwater resources cannot be directly based
on fixed observation points. It requires taking into account the groundwater flow for
large areas and long periods of time, involving a great deal of uncertainty. In such a
situation it is especially difficult to reach a common approach between the negotiating
parties.
The history of the Lebanese-Syrian Agreement is a good example of how
groundwater flow has eventually been taken in account. It also makes clear how far
groundwater diversity is taken into account and remains partly absent because of
the difficulty of determining simple and effective rules based on observations on the
subsurface environment.
The creation of new numerical procedures for water flow simulation provides a
useful aid for the negotiators. The outcomes nevertheless depend on robust data and
the establishment of a coherent conceptual model, with which underpin the basis for
the negotiations.
REFERENCES
Comair G.F., McKinney D.C., Scoullos M.J., Flinker R.H., Espinoza G.E. (2013)Transboundary
cooperation in international basins: Clarification and experiences from the Orontes River
Basin agreement: Part 1. Environmental Science & Policy 31, 133140.
Comair G.F., McKinney D.C., Scoullos M.J., Flinker R.H., Espinoza G.E. (2013)Transboundary
cooperation in international basins: Clarification and experiences from the Orontes River
Basin agreement: Part 2. Environmental Science & Policy 31, 141148.
Droubi A. (2012) Communication interne. Meeting du projet IHEID Orontes River basin
Liban-Syrie. Novembre Bekaa, Lebanon.
ESCWA13 (2013) Inventory of shared water resources in Western Asia. Chapter 7, Orontes
River Basin, UN, ESCWA/SDPD/2013/WG13/REPORT.
FAO (2011) Aquastat database. Syria Arab Republic. http:www.fao.org/nr/water/aquastat/
countries/syria/index.stm; last visit 10.01.2012.
Hamade S., Tabet, Ch. (2013) The impacts of climate change and human activities on water
resources availability in the Orontes watershed: Case of the Ghab Region in Syria. Journal
of Water Sustainability 3(1), 4549.
Kloosterman F.H., Vermooten, J.S.A. (2008) Final Report, development of a numerical ground-
water flow model for the Larger Orontes Basin. Dutch-Syrian Water Cooperation, TNO
report.
Kloosterman F.H. (2009) Notes on the Origin of the Groundwater issuing from The Ain
Altnour, Ain Asamak and Ain Alzzarka Springs. Mission Report EVD Syria Bridging Phase.
TNO/Deltares.
Maalouf F. (1999) Approche du fonctionnement de laquifre de lOronte. DEA Univ. St Joseph,
Beyrouth.
Sarraf S. (2013) Connecting transboundary water resources management with national visions
and plans in Lebanon. Near East & North Africa Land and Water Days, Amman, 19.
This page intentionally left blank
Chapter 5
HungarianSlovakian
transboundary karstic
groundwater management
under the scope of ENWAT and
TRANSENERGY EU projects
Peter Malk 1, Radovan C ernk 1 & Gyrgy Tth 2
1
ttny geologick stav Dionza tra Geological Survey of Slovak Republic ,
Bratislava, Slovakia
2
Geological and Geophysical Institute of Hungary, Budapest, Hungary
ABSTRACT
Within the framework of European Union projects ENWAT and TRANSENERGY, Hungarian
and Slovakian hydrogeologists cooperated on common descriptions, modelling and formu-
lating proposals for the groundwater management of transboundary karstic aquifers, one
of them a deep-seated geothermal aquifer and the other a classical plateau karst structure.
Some notes from lessons learned during the work on both projects may be useful to other
hydrogeologists trying to establish cross-border cooperation especially in Europe, such as
maintaining two independent web portals for public use and for specialists. Keeping units to
be constantly linked with the data in common databases of input parameters, and adequate
(possibly metric) coordinate systems seems to be useful. Choice of software tools should
be left for individual selection by working teams, while basic standards of data inputs and
outputs seem to be the most effective common understanding for hydrogeologists managing
transboundary aquifers.
5.1 INTRODUCTION
GERMANY
POLAND
CZECH
REPUBLIC
SLOVAKIA
UKRAINE
2
AUSTRIA
1
HUNGARY
ITALY SLOVENIA
ROMANIA
CROATIA
Figure 5.1 Location of the two discussed transboundary aquifers between Hungary and Slovakia
(1:Komrom-trovo geothermal karstic body, 2:Aggtelek/Slovensk kras transbound-
ary aquifer).
2008; Malk et al., 2012; No.2 on Figure5.1). Another project, dealing with trans-
boundary groundwater is TRANSENERGY (Transboundary Geothermal Energy
Resources of Slovenia, Austria, Hungary and Slovakia). The project was targeted
at geothermal waters (geothermal resources) in several regions: the pilot area of
Komrom-trovo (Komrno-Prkny) is a deeply seated geothermal karstic aquifer
(Ndor et al., 2013; No. 1 on Figure5.1).
One of the three pilot areas of the ENWAT project was the Aggtelek Mountains and
the Slovensk kras Mountains (Slovak Karst). The area represents a large common
karstic aquifer system in the Eastern part of both countries (Figure5.2). Both coun-
tries considered it as a highly important transboundary water body. The area is pro-
tected as a National Park on both sides of the border, the Aggtelek National Park
(Aggteleki Nemzeti Park) in Hungary and National Park of Slovak Karst (Nrodn
park Slovensk kras) in Slovakia. In 1995, all the caves of the Aggtelek and Slovak
Karst were designated World Heritage Sites by UNESCO. Significant drinking water
resources in Slovakia and regionally important resources in Hungary, although still
preserving sufficient quality, are located in vulnerable karstic area and require preven-
tive protection.
The groundwater body is in a Mesozoic complex with morphologically visible
karstic plateau and canyon-like river valleys, separating different hydrogeological
units. Outcropping rocks are very different according to the character of permea-
bility, character of groundwater circulation, type of groundwater regime, and also
in the resulting yield of springs. From the hydrogeological point of view, the most
HungarianSlovakian transboundary karstic groundwater management 65
important tectonic unit in the area is the Silicikum unit, mainly its Middle Triassic
and Upper Triassic part. The most important aquifer with karst-fissure type of per-
meability is formed by the Middle and Upper Triassic limestones and dolomites.
Similarly there are important hydrogeological units on the Hungarian side such as
Alshegy, Nagyoldal, Haragistya and Galyasg, which contain the Aggtelek-Domica
cave system. Tertiary filling of sedimentary basins (mostly clays and clayey sands)
act as aregional impermeable barrier for the groundwater accumulated in Triassic
carbonates.
Natural groundwater circulation in karstic aquifers is typified by intensive drain-
age towards the major springs. Their infiltration areas are not easy to delineate, but
results of the water balance suggest that at many places groundwater infiltrating in
one country is draining towards a spring located in the neighboring country. For
example, the underground hydrological system of the Domica-Baradla cave system
is fed by water recharged in Slovakia, while the major drainage element here is the
Jsva spring at Jsvafo in Hungary. Another example is the Alshegy/Doln vrch
karstic aquifer unit, elongated along an east to west axis, with a national boundary
crossing over it. The eastern part of this aquifer drains towards springs in Hungary,
while natural groundwater outlets are not present on the Slovak territory. By contrast,
66 Karst without Boundaries
the western part of the Alshegy/Doln vrch karstic aquifer drains towards north, to
springs located on the Slovak territory.
Groundwater chemical composition or quality originates mainly from water-
rock interaction. Groundwater circulates predominantly in limestones and dolo-
mites of the Mesozoic rock formations of Middle and Upper Triassic age. The
Ca-HCO3 and Ca-Mg-HCO3 types of chemical content dominate in groundwater
that is not affected by anthropogenic activities. The natural character of ground-
water circulation is at the moment still able to preserve good properties of high
quality groundwater and most of the area provides high quality groundwater for
regional waterworks. This groundwater fully meets the criteria for drinking water.
In most cases, concentrations of trace elements are low and mostly below the limits
of detection. Anthropogenic contamination was found only rarely, usually close to
settlements. It was indicated by raised concentrations of nitrate, chloride, sulphate,
chemical oxygen demand and potassium. High iron and manganese concentrations
may be found at the adjoining Tertiary artesian aquifers which have a reducing
environment.
In the Aggtelek/Slovak Karst area the chemical status of groundwater is good, but
possible future climate change may affect this with increased occurrence of extreme
hydrological events. In the worst scenario, from the water management point of view,
higher and more rapid flood peaks will apply pressures to water channels by ero-
sion while drought periods damage the ecology of the fragile karst area. Small-scale
water management arrangements buffering hydrological extremes, such as building
of minor water storage systems enable water supply to continue in dry periods, and
polders, dry reservoirs to collect high water stages are the best strategy. A significant
threat in this economically less developed part of Slovakia and Hungary is uncon-
trolled land use and building on flood-prone areas. [More details are found on the
project output website http://www.all-in.sk/enwat, maintained in Hungarian, Slovak
and English languages.]
geothermal resources and sustainable reservoir management, gained during the proj-
ect, was transferred to end-users to provide an overview and for them to make simple
estimations on geothermal reserves within the project area. This publicly accessible
implementation tool shows all the relevant information on the potential, vulnerability
and sustainability of the geothermal system in the transboundary regions.
Results integrate all of the activities carried out within the TRANSENERGY
project, such as screening utilisation needs with special respect to national, EU and
international legislation, collecting and harmonising geoscientific data, performing
additional measurements, organising all data in harmonised, multi-lingual joint data-
bases, producing various cross-border geoscientific models and performing scenario
modelling for different extraction of geothermal heat/water.
In addition to the web-based decision planning tool, the project also delivered
a methodology for joint groundwater management and utilisation maps summaris-
ing the legal steps and actions towards a harmonised management strategy of trans-
boundary geothermal resources, and a best practice on geothermal use. The project
partners were the Geological Surveys of the four countries Hungary, Slovakia,
Austria, Slovenia, which are involved in several international projects concerning nat-
ural resources. They also have a long tradition of bilateral cooperation with each
other in several projects.
Five pilot areas were treated within the TRANSENERGY project, while one
of them the Komrom-trovo (Komrno-Prkny) pilot area is a deeply seated
karstic groundwater body (Figure5.3). The Komrom-trovo pilot area belongs to
the Komrno block, in Slovakia Komrno high block and Komrno marginal block
Figure5.3 Position of the transboundary Komrom-trovo karstic geothermal aquifer (after Gspr
and Tth 2013).
68 Karst without Boundaries
are distinguished (Remk et al., 1992) comprising a subsided northern blocks of the
Gerecse and Pilis Mountains (Hungary).
The main and most important aquifers in the pilot area are the Upper Triassic plat-
form limestones and dolomites (Dachstein Limestone and Main or Haupt Dolomite).
The Middle Eocene denudation caused strong karstification in the more than 1500
m thick carbonate sequence. These well karstified conduits and fractures along the
tectonic elements determine the karstic groundwater flow; the karstification in the
upper part of the system has higher permeability and this is where the main part of
the groundwater flow takes place. From the area of outcropping Upper Triassic rocks
(North-Bakony, Vrtes, Gerecse, Pilis mountains in Hungary) the recharged precipita-
tion descends and flows towards the deeper regions to the north west and west. From
the north west edge of the aquifer the water turns towards north-north east and in the
Slovakian parts towards the east.
Along the margin of the mountains (Tata, Dunaalms, Patince and Esztergom)
natural discharge areas of the lukewarm karst springs (~2027C) discharge in:
Esztergom springs at 2629C, Tata springs at 2022C, Dunaalms-Patince sprigs
higher at 2324C and 2527C. The higher (~40C) water temperatures in trovo
(against ~28C in Esztergom) can be explained by longer flow paths from the north
west and west (Gspr & Tth, 2013). The marginal (west, north west and north)
and deeper part (more than 600 m below the sea level) of the Upper Triassic carbonate
aquifer is characterised by higher temperatures and belongs to the Komrno marginal
block, is without hydraulic contact with the Komrno elevated block and has no
recharge area (Remk et al., 1992). This thermal karst water (4060C) is produced
by deep wells in the north west and north of the area (near Bbolna, Komrom and
Komrno).
The main utilisation is for health purposes in both countries in the north east part
of the area (Esztergom and trovo). In Patince and Dunaalms (historical) balneo-
logical and drinking water uses exist. Near Komrom and Komrno balneological
and agricultural use prevails. Most of the users utilize the lukewarm or thermal water
of the Triassic karstic aquifer, but some Miocene and Cretaceous local aquifers near
Komrom and Komrno are also exploited.
Because of the long-term brown coal mining coupled with intensive water abstrac-
tion (Tatabnya Dorog area), the whole cold- and thermal karst system was affected
by a regional depression which caused the drying out of most of the lukewarm springs.
After the mining ceased the water level has been rising since the beginning of the
1990s (Alfldi & Kapolyi, 2007). In this dynamically changing system it is hard to
estimate the actual dimension of the impacted area in karst (difference between natu-
ral groundwater level in karstic aquifer existing prior to mining activities and its cur-
rent state), but in the south and south west part of the area in 2013 it was about 30m,
and decreases toward the norh. Along the Danube river from Komrom-Komrno to
Esztergom-trovo the actual drop in the water level remaining as a consequence of
the past mining activities in the karst was about 10m. Existing abstraction for drink-
ing water supply in the area is much smaller than the mining abstraction in the past
and has a potentially low influence on the main lukewarm springs. In the area of Tata
the rising karst water level results in seepages on the surface and affects surface infra-
structure such as buildings and garages. Therefore, solutions have to be found to uti-
lize and/or drain the (surplus) seepage waters. In this area there is also a competition
HungarianSlovakian transboundary karstic groundwater management 69
between the water demand of balneological users, the drinking water abstractions and
protection of the groundwater dependent ecosystems, which represents a very import-
ant environmental aspect with high priority in the EU Water Framework Directive.
Therefore, ranking of different needs and an integrated assessment of their impacts
is of vital importance. The coordinated utilisation and planning of further develop-
ment is needed in both countries to maintain the current situation and the historical
heritage. In the area of the natural discharges (Tata, Esztergom in Hungary, Patince
in Slovakia) the thermal pollution of the surface waters is minimal and nature has
adapted to lukewarm karst waters.
The assessment of sustainability was based on an overview of 34 geothermal
wells, 8 on the Slovakian side and 26 on Hungarian side (in the Komrno high block).
No over-exploitation was reported. Due to abandoning mining areas and halting of
water pumping, water levels are rising. However, some negative effects on infrastruc-
ture were reported and a continuous karstic water level monitoring programme was
proposed.
The hydrogeological structure of the Komrno marginal block that is closed to
water circulation and has no hydraulic contact with the surrounding structures does
not have reinjection of geothermal water (Gspr & Tth, 2013, vasta et al., 2013).
Although water has no borders and flows freely between the countries, the cat-
egorisation and delineation of the water bodies is not harmonised across the border.
Both countries have a different way and philosophy of dealing with the delineation of
geothermal water bodies. For a mutual assessment and management of the ground
water bodies, harmonised rules are needed in the future.
Results of both projects, based on evaluations and models, local needs, costs and best
practices, are a step forward in the creation of a joint HungarianSlovakian water
management plan by supplying basic data and new information on transboundary
groundwater bodies. The work also indicated the need to continue the process of
potential pollutant characterisation, risk assessment and common regulations to con-
trol water abstraction in both countries. Some notes from lessons learned during the
work on both the projects may be useful to other hydrogeologists trying to establish
cross-border cooperation especially in Europe. (a) Maintaining two independent web
portals has proved to be useful to keep the project agenda for both specialists (work-
ing portal as a place of exchange information during project between working groups)
and public (public portal to publish preliminary and final results of the projects via
lay reports, as well as by more technical and scientific outputs). (b) Common database
of input parameters that should allow simple manipulation and visualisation of data
and contain forms to browse and edit the data: the role of showing units for every
component as frequently as possible is inevitable! Majority of misunderstandings and
time consuming discussions and explanations were created by different units used
by respective working groups. (c) Coordinate system: practically all countries keep
the geospatial data in their national (and usually mutually incompatible) systems;
the Hungarian EOV system and Slovakian/former Czechoslovakian S-JTSK Krovk
system. Popular and frequently used Longitude/Latitude WGS84 shows distances in
70 Karst without Boundaries
REFERENCES
ABSTRACT
The West Stara Planina Mountains are situated across two countries Serbia and Bulgaria. The
karstification is developed mainly in two carbonate complexes: Triassic and Upper-Jurassic,
while in the western extension, i.e. in Vidlic Mountain, karstified carbonate rocks of Cretaceous
age prevail. This is a typical mountain karst with wide distribution of classical karstic land-
forms: dolines, small poljes, and blind valleys. There are also hundreds of caves in both coun-
tries, many of them well explored. Due to high infiltration of rainfall, abundant groundwater
reserves are formed. The drainage takes place over numerous large karst springs. There are
overflow sources while other sources are from upwelling water, draining the deeper saturated
zones. Their discharge regime varies from relatively constant to highly changeable. Some are
tapped and used for potable water supply, as in the case of Pirot, Dimitrovgrad (Serbia) or
Svoge (Bulgaria). Most of the karst springs have relatively clearly defined inner catchment areas
and are not subject to transboundary discharge. The water budget analysis at the bordering ter-
ritories of Serbia and Bulgaria requires more hydrological, hydrogeological and climatic data,
field survey and water tracings.
6.1 INTRODUCTION
The Balkan Mountains (Stara Planina Mountains) are a mountain range in the eastern
part of the Balkan Peninsula and represent a part of the Alpine-Himalayan orogenic
belt. The western part, located on the territory of Bulgaria and Serbia, has, for various
reasons, not been well explored. In the absence of a common survey and exact data
there is considerable speculation concerning water flows from one side of the border
to the other. During recent years, the joint research of Serbian and Bulgarian scien-
tists, both sponsored by their Academies of Sciences in the fields of geology, geomor-
phology and hydrogeology, has been active. One of their most interesting and topical
problems having an important ecological and practical significance was to standardise
the research for cross-border areas to solve the problems of karst aquifer distribution
and karst water flow directions.
72 Karst without Boundaries
The target of the study is the part of the Western Balkan Mountains where karstified
rocks are continuously distributed throughout the territories of the two countries.
Such areas are located south of the main ridge of the mountain, between the Temska
River (in Serbia) and the Iskar River (in Bulgaria) (Figure6.1). The southern bound-
ary of the area is marked by rivers draining surface and underground waters of this
part of West Stara Planina the Niava River in Serbia and the Elovitsa and Blato
Rivers in Bulgaria. The total surface of the transboundary region is about 2000 km2.
The landscape is typically mountainous. The highest elevations between 1300
and 2015 m a.s.l. (peak of Kom) are on the main ridge of Stara Planina. South of
and parallel to the main ridge are series of ridges and valleys. The lowest parts are
located along the southern border from about 300 to 700 m a.s.l. The relief pro-
vides various climatic conditions. The annual average air temperature fluctuates from
3.4C to more than 10C according to the altitude. The annual rainfall varies widely
from about 300 to 1000 mm (Koleva & Peneva, 1990; Stevanovic, 1991; Ristic,
2007). According to the graph based on data from the rainfall stations in Bulgaria and
Serbia, there is a good correlation between rainfall and altitude (Figure6.2).
The formation of karst and karst aquifers depends mainly on the geological condi-
tions. They have been studied by several researchers and the results are summarised in
geological maps at scale 1:100000, map sheet Pirot (Geological Survey Serbia, 1970)
on Serbian territory and map sheets Pirot (Haidutov, 1992), Berkovitsa (Haidutov &
Dimitrova, 1992) Vlasotince & Breznik (Zagorchev & Kostadinov, 1991) and Sofia
(Yanev, 1992) on Bulgarian territory. Between the sheets on both sides of the state
border there are some discrepancies in the geological boundaries and rock outcrops
r
ive
kaR
ms Ma SERBIA
Te in
rid
ge BULGARIA
of
Sta
ra
Pla
nin
a
BULGARIA
er
SERBIA
Iskar Riv
Ni
sh
av
a
Riv
er
1 4
2 5
Bla
to
3 6 0 10 20 River
kilometers
Figure 6.1 Geological sketch map of West Stara Planina Mountains. 1 Upper Jurassic Lower
Cretaceous limestones; 2 Triassic limestones and dolomites; 3 Main faults; 4 State
boundary; 5 Main ridge of Stara Planina; 6 Rivers.
Development and protection of transboundary karst and karst aquifers 73
1600
Y = 1,2108x 98,507
1400 R2 = 0,9163
1200
Altitude, m
1000
800
600
Serbia
400
Bulgaria
200
0
0 500 1000 1500
Precipitation, mm
Figure 6.2 Correlation between rainfall and altitude in Western Stara Planina.
BULGARIA
SERBIA
1 5
2 6
3 7
A B C
4 a
8 0 10 20
b
c kilometers
Figure 6.3Karst in West Stara Planina Mountains. 1 Karstified rocks of Southern zone; 2 Karstified
rocks of Northern zone; 3 State boundary; 4 Rivers; 5 Lakes; 6 Poljes; 7 Blind
valleys; 8 Caves (Length: A >3000 m, B 10003000 m, C <1000 m; Height differ-
ence: a >100 m, b 50100 m, c <50 m).
higher parts of the mountain, the two carbonate complexes build monoclinal struc-
tures dipping to the south and separated by non-karstic LowerMiddle Jurassic rocks.
To the south it is bordered by Lower Triassic sandstones and Paleozoic rocks in the
higher parts of the mountains. The southern boundary is a thrust belt of east-west
direction passing through the study area.
The second, the Southern zone, is characterised by more complex block structures
due to a series of horst and graben structures also oriented east-west. The complex
tectonics are reflected on the landscape of this zone with a series of depressions
linearly oriented in the same direction separated by ridges.
The river network also has an important impact on the karst and karst waters.
Rivers are usually formed in the highest parts of the mountain and flow to the south.
Some of them (mainly on the territory of Bulgaria) completely lost their flow entering
the karst terrains. Some of the rivers cross the limestone and dolomite in the first zone
and flow periodically after rainfall: for example the Gintsi River, the upper reaches of
the Visochka (Visocica) River, and the upper reaches of its right tributaries in Serbia
(Rosomacka, Jelovicka). The Visocica and Niava Rivers represent the major drain-
age arteries of the second (southern) zone and the Iskar River drains only the eastern
part of the northern zone. Several rivers such as the Iskrets and Blato in Bulgaria begin
from large karst springs. The western boundary of the area the Temska River has
almost no connection with karst.
The Niava River crosses the border between Bulgaria and Serbia near
Dimitrovgrad with a total length of 202 km to its confluence with Juna Morava
River and a total catchment area of 4068 km2 (Dukic, 1975). About 70% of the
region falls within the catchment area of its upper reaches, while the remaining east-
ern part of the region is in the catchment area of the Iskar River and its tributaries.
The Niava originated from the higher parts of the Stara Planina Mountains from
numerous springs. In the upper part of the watercourse it is called River Ginska.
Entering karst carbonate rocks, it lost a substantial part of its flow only to reappear
in Godech where it had already been named Niava.
The Visocica River is also a cross-border river, which after its confluence with
the Temska River flows into the Niava River. This river also originates from the high
parts of the Stara Planina in Bulgaria and after entering Serbia is a major drainage
artery of the northern, high part of the study area. Only a small tributary south of
Visocica River passes from Serbia to Bulgaria and flows into Niava River. The char-
acteristics of the quantitative parameters of the Niava and Visocica Rivers, before
their entry into Serbia, are described in detail from Hristova (2010), as the average
annual water quantities of the Niava vary from 0.91 to 3.47 m3/s. According to
Dukic (1975) the surface runoff for the territory of Bulgaria at Dimitrovgrad station
amounted to 5.47 l/s/km2, while in Pirot, where the Niava River flow out of the
investigated region is 6.63 l/s/km2.
About 60% of the total area of the Western Balkans is occupied by karstified rocks.
The first mention of the distribution of karst landforms in the West Stara Planina
Mountans was by the Austrian geologist Franz Toula (1882). Toula reported dolines
Development and protection of transboundary karst and karst aquifers 75
and karren fields in the area between Vidlic Mtn and Niava River (Toula, 1882;
Radev, 1915).
Some notes about karst landforms in the studied area are published by Jovan
Cvijic in his first karstological monograph Caves and Underground Hydrography in
Eastern Serbia (1895) (Calic, 2007). The karst relief is described in the book Karst
Landforms in West Stara Planina (1915) by the Bulgarian geomorphologist Zheko
Radev. Radev surveyed 11 caves and performed a detailed study of the surface karst
landforms: poljes, uvalas, dolines and karren fields, supported with photos (Radev,
1915).
During the last several decades the karst geomorphology and speleology in West
Stara Planina have been described by several authors: Petrovic (1974, 1976), Benderev
(1989), Angelova et al. (1995, 1999), Beron et al. (2006), Zlatkova (2006), Kostov
(2008), Szeidovitz et al. (2008), Mihaylova et al. (2008, 2009), Shanov & Kostov
(2015).
From a morphological point of view the karst is an exposed, mountain type
with widespread surface karst landforms. The intensity of karst processes is differ-
ent in the Northern and Southern zones of the studied area. In the Northern zone
the karrenfelds, dolines and uvalas are widespread. The significant blind valleys
play an important role especially on the territory of Bulgaria. Compared to the
Southern zone, a large number of caves and potholes, for instance 184 in Bulgaria,
are established here. The caves are widely distributed and the morphology has deter-
mined the monoclinal dipping to the south of the layers of carbonate rocks with
old or active ponors of surface water. There are also caves representing active and
temporary springs (Vodnata Cave, Krivata Cave, Dushnika Cave, etc.). In this zone
most of the deepest and longest caves are located in the Western Balkans (Table6.1,
Figure6.3).
Table 6.1 The longest and the deepest caves in the studied area.
The karst distribution in the Southern zone varies in the different areas, depend-
ing on the specific geomorphological and hydrogeological conditions. There are both
exposed karst areas and sections covered with thick soil layers. The distribution of
uvalas and dolines is different: there are both areas with significant dolines density (for
example 164 dolines on the flattened ridge part of Tri Ushi Ridge in Bulgaria over an
area of 35 km2 Figure6.4) and areas with relatively low levels of karstification. Typical
of the zone are the large karst poljes: Rayanovsko, Dragomansko and Aldomirovsko
poljes in Bulgaria, and Odorovacko in Serbia (Figure6.5). According to the polje clas-
sification scheme of Gams (1994), the poljes in the Bulgarian part are of the overflow
type. In Dragomansko and Aldomirovsko poljes permanent marshes are formed.
The number of caves in the Southern zone is much lower, only about 60 in the
Bulgarian part. Most of the caves are small being up to 100 m long and 25 m deep.
The exceptions are the Vetrena dupka Cave in Serbia (4150 m), and the Golyamata
Temnota Cave (2100 m) and Temnata Dupka Cave (493 m) in Bulgaria. The latter
two are active ponor caves.
Figure 6.4 Karst dolines on the flattened ridge part of Tri Ushi Ridge in Bulgaria.
The data on the karst aquifers and waters in the area are summarised separately for
Bulgaria (Antonov, 1963; Antonov & Danchev, 1980; Benderev et al., 2005) and
Serbia (Stevanovic, 1991, 1994; ivanovic, 2011). Along with the karst waters in the
transboundary region, there are both fracture and intergranular aquifers (Figure6.6).
Fracture aquifers are formed in different aged rocks, with different lithological and
petrographic features. Tectonic conditions determine complex relationships between
karstic rocks and the other pre-Quaternary massive rocks. In the North, these rocks
are relatively poorly permeable and underlie the karst aquifers. In the South, because
of the block tectonics, the karst and fracture aquifers are characterised by different
permeability, and their recharge depends on contacts (lithological and tectonic) with
impervious rocks. Intergranular aquifers are formed in unconsolidated Quaternary
and Pliocene deposits. Pliocene and Quaternary sediments in the southern part of the
area have more significant hydrogeological importance and form independent hydro-
logical units. In their underlying formations blocks of karst rocks, which outcrop in
the uplands, and have hydraulic connection with them, are included.
The karst aquifer is primarily recharged by rainfall. The rainfall quantity increases
from south to north (Figure6.2), due to high precipitation and the converse reduction
of the evapotranspiration in the higher parts of the mountains. In the northern zone
an important component of the water balance is temporary and permanent infiltration
BULGARIA
SERBIA
1 5
2 6
3 7
a b 0 10 20
4 8
kilometers
Figure 6.6 Hydrogeological map of West Stara Planina Mountains. 1 Karstified rocks (Southern zone);
2 Karstified rocks (Northern zone); 3 Rock with fissure water; 4 Quaternary and
Pliocene sediments with porous groundwater; 5 Lakes; 6 Rivers; 7 State boundary;
8 Main spring (Temperature: a T < 15C, b T > 15C).
78 Karst without Boundaries
of surface rivers descends from the main ridge of Stara Planina. In the southern part
the infiltration rate of surface water is much lower. It is related to those watercourses
passing through karst poljes.
According to the water budget calculations by Stevanovic (1994) for the
Serbian part of Stara Planina Mountain, 18% of the precipitation recharges the
karst groundwater. Water balance calculations for Bulgaria were made only at two
isolated areas. The first occupies the eastern part of the northern area of the distri-
bution of carbonate rocks (east of Ginska River) called Ponor Mountain (Benderev,
1989). The altitude for this area varies from 500 to 1400 m while the rate of evapo-
transpiration varies from 70 to 36% of rainfall (Turc method). Typical of this region
is that almost the entire surface runoff recharges the karst groundwater and thus
forms around 20% of the water reserves (regime data of the measurement of river
quantity). The second area coincides with the catchment areas of the most southerly
situated springs Opitsvet, Bistrets and Bezden, located at elevations between 550
and 900 m. Conducted studies used water balance methods showed that 22% of
the precipitation recharges the groundwater system (Mihaylova & Benderev, 2010).
This value is relatively close to the average for the Serbian territory where low alti-
tudes prevail.
The general direction of groundwater movement is from north to south, flowing
to the lowest parts of the terrains. One exception is in the easternmost part of the
region, where the karst water is directed to the Iskar River.
The drainage is mostly by springs (Table6.2, Figure6.6), by groundwater
abstraction and by underground flows into lateral permeable intergranular aquifers.
There are both small gravity springs of local catchment areas with low flow rates and
typical ascending karst springs with high flows. In the northern zone are springs with
relatively high and variable discharges including Peshta (Iskrets spring), Jelovicko
vrelo and Skaklya. These springs are characterised by conduit flow and very fast
response to precipitation events (Eftimi & Benderev, 2007). Springs of this type exist
also in the southern zone at Bistrets and Berende izvor. Fault tectonics associated
with the rise of some and the sinking of other blocks in many places create conditions
for the formation of saturated zones which are drained by springs with more stable
flow rates Staroplaninsko oko, Kavak, Ropotski. Interesting are the larger drainage
areas where there are several springs draining different hydrodynamic zones on the
border of Stara Planina with Sofia and Pirot plains; for example Krupac 1 and 2,
Topli izvor in Serbia and Bistrets, Bezden, Opitsvet in Bulgaria. In these areas both
springs drain the upper part of the saturated zone with highly variable flow and the
springs at lower elevations have a smaller range of flow rates. Characteristic is the
presence of ascending springs with higher temperatures (1922C) draining areas
with deep circulation the Topli izvor and Toplo vrelo Springs in Serbia and the
Opitsvet in Bulgaria.
The specific characteristics of the hydrogeological conditions and the existing infor-
mation for the transboundary region determine the approach used for the evaluation
of groundwater vulnerability to pollution. The need to use the karst waters leads to
Development and protection of transboundary karst and karst aquifers 79
Discharge, l/s
ecological problems because of the water quality. One of the most important problems
in the area is the correct recognition of the sites of potential pollutants to the karst
waters, as well as the need for zones of increased protection.
The EPIK method is a GIS-based, multi parametric method which takes four
parameters into consideration: Epikarst, Protective cover, Infiltration condition and
Karst network development (Doerfliger et al., 1999; Zwahlen, 2003). It has been
chosen as the most suitable approach for describing the vulnerability of the karst
areas. It was created especially for areas of mountain karst and has been successfully
applied for different karst basins (Gogu & Dassargues, 2000; Iurkiewicz et al., 2005;
Goldscheider, 2005; Awawdeh & Nawafleh, 2008). Maps have been drawn sepa-
rately for the adjacent territories of Serbia and Bulgaria (ivanovic & Dragiic, 2013;
Mihaylova et al., 2009), and they are unified and adjusted in the present study.
The rates of significance of the different EPIK parameters are determined by their
weight coefficients. Each of these parameters, taking into account its peculiarities, is
80 Karst without Boundaries
divided into a number of classes. These classes consider the quantitative and the qual-
itative indicators impacting the protection of the groundwater.
The weight coefficients of the parameters and the numerical values of the classes
have been argued and determined in Doerfliger et al. (1999); Zwahlen (2003). Data
for EPIK Vulnerability Index are categorised into four groups Very high, High,
Intermediate and Low Vulnerability. They can be visualised on GIS prepared maps.
Figure6.7 presents the terrains of Very high and High Vulnerability for the Stara
Planina Mountain region in Bulgaria and Serbia. The zone of Very High Vulnerability
covers 157 km2, or about 7.5% of the whole area, while the zone of High Vulnerability
occupies 605 km2 or 29% of the whole.
The distribution of the zones of Very High and High Vulnerability has significant
importance for protection and management of the karst groundwaters. The zones
are situated in areas of relatively less important anthropogenic impact. The most sig-
nificant urban areas are the towns of Pirot, Dimitrovgrad, Dragoman and Slivnitsa,
situated on low relief parts at the periphery of the region, and outside the karstic areas
including the zone of intensive groundwater recharge. Industrial units exist only in the
town areas. The towns and the villages are connected by roads but the overall traffic
is low.
The Neogene sediments filling one of the east-west oriented grabens across the
national border contain lignite beds. Excavation of these coals is carried out near
the village of Staniantsi in Bulgaria and in the past also at the village of Mazgosh in
Serbia. These two areas are situated outside the established zones of Very High and
High Vulnerability, and they do not impact directly the karst waters.
BULGARIA
SERBIA
1 5
2 6
3 7
0 10 20
4
kilometers
Figure 6.7 Zones of Very High and High Vulnerability. 1 Very High Vulnerability; 2 High Vulnerability;
3 State boundary; 4 Rivers; 5 Lakes; 6 Towns; 7 Villages.
Development and protection of transboundary karst and karst aquifers 81
The town of Godech is situated in the centre of the studied area in Bulgaria, but
it is also built on non-karstic rocks, on the bank of the Niava River. Most of the
other less significant inhabited sites are also situated outside the outcrops of karsti-
fied rocks, predominantly on terrains used for agriculture. A limited number of small
villages, situated mainly in the lower parts of the river valleys, are within the zones of
Very High and High Vulnerability.
There are a few villages situated in the zones of outcropping karst, where karst
aquifers are recharged. An example is the village of Ponor in Bulgaria, built inside an
large surface karst form. On the territory of Serbia few villages are situated on large
karst poljes.
The inhabited sites are surrounded normally by cultivated terrains, and animal
breeding is also important for the local population. Agricultural cultivation in the
zones where karstified rocks exist is possible only where uninterrupted soil cover
on these rocks is present. Such conditions exist on the flattened bottoms of the karst
forms, as karst poljes or ponors. These cultivated terrains are one factor for the
increase in karst water pollution.
Many quarries exist both in Bulgaria and Serbia where limestone and dolomites
from the vadose zone are exploited. It is commonly accepted that they do not affect
karst waters, even though some of them are close to the discharge zones and karst
springs. One example is Iskrets Karst Springs which have a variable discharge from
280 l/s to more than 50000 l/s (Benderev, 1989). They drained more than 80% of the
territory of Ponor Mountain, a part of Stara Planina Mountain inside the area studied
in Bulgaria. Precipitation over the capture area of the springs (about 140 km2) pro-
vides 62% of the average annual discharge of the springs.
The springs have dried up several times during the twentieth century. After the
Vranchea Earthquake in 1977 (M=7.2) and at a distance of about 400 km from
the site, the discharge rate dropped from 5.5 to 500 l/s in 7.5 hours (Paskalev etal.,
1992). Later, the discharge rate rose abruptly to 13500 l/s, and then began to decrease
gradually. Similar events happened during the local Svoge Earthquake of 1979, as well
as on 11 April 1982, when no significant earthquake was recorded at the localities or
in the whole Balkan Region.
The question of the seismic impact is of importance for the local authorities
because of the significance of the springs as a major source of fresh water for the
town of Svoge. The quantities normally used are at the rate of 150 l/s. It was proved
that the blasts in the quarries at the nearest vicinity of the springs do not have the
potential to disturb the normal discharge of the springs (Shanov & Benderev, 2005).
The only factor for such disturbances can be the local tectonics and the dynamics of
the processes inside the karst system (Shanov & Kostov, 2015).
The most important hydro-technical facility built on the Bulgarian side is a canal
designed for redirecting a part of surface waters, which recharges naturally in high
parts of the mountains, towards the neighboring watershed for their use for hydro-
electric power generation.
The most important hydro-technical facility in Serbia is Zavojsko Lake which
dams the Visocica River. The biggest part of the lake is outside the area of karstified
rocks, except for a limited zone near the dam. This lake was formed in 1963 after
a spectacular landslide on the right river bank which dammed the Visocica River. A
50 m high natural dam was formed and caused total flooding of several upstream
82 Karst without Boundaries
villages. The landslide activated as an effect of abundant water infiltration from snow
melt water. Recently this dam was replaced by an artificial one providing the power
for a hydro-electric power plant.
Karst springs on both sides of the border are an essential source of water supply
for the settlements. Karst waters are tapped mostly at the springs with gravity trans-
port towards consumers. Most of the existing springs are used, as the large water
quantities are taken to supply the larger settlements, such as Krupac springs for Pirot,
Vrelo Protopopinci for Dimitrovgrad, Opitsvet for Slivnitsa, Zli dol for Godech,
Peshta (Iskrets spring) for the town of Svoge. However, karst groundwater is not
directly affected by the exploitation and there are sufficient amounts of water for
dependent eco-systems. In contrast, significant impact on the balance of karst waters
may be caused by exploitation wells drilled in karst aquifers covered by Quaternary-
Pliocene deposits in the Sofia and Pirot basins. At this stage that impact is negligible
because of the small number of wells using water from karst aquifer and small rates
of abstraction (Directorate for the Danube River Basin, 2015).
Northern zone
SERBIA BULGARIA
1300 m
Komshtitsa r.
1200 m
Komenicka r.
1100 m
1000 m
900 m
2.5 km 5.0 km 7.5 km 10.0 km 12.46 km
Southern zone
750 m SERBIA BULGARIA
700 m
Nishava
650 m
600 m
Figure 6.8 Schematic terrain elevation profile of transboundary areas in the Northern and Southern
zones.
Development and protection of transboundary karst and karst aquifers 83
to the river about 34 m above it (Beron et al., 2006) and where the river waters enter.
River waters percolate down to the saturated zone but because of a fault against the
elevated south block which passes through Serbian territory, the water probably flows
to the west, where the lowest possible drainage areas are.
In the Southern zone the main drainage artery is the Niava River but the probable
influence of the transboundary impact is minimal due to the small catchment areas,
absence of a significant river recharge into the area and lower level of karstification.
The zone represents a low mountain hill with leveled ridge areas with karst for-
mations formed on both Bulgarian and Serbian territory. Precipitation can be drained
away in different directions; to Berende spring in Bulgaria and to the springs in
Dimitrovgrad and Krupac in Serbia, but there is no information on the spatial posi-
tion of the underground watershed.
Neither of the transboundary regions are considerably affected by human
activities.
6.7 CONCLUSIONS
The preliminary analysis of the existing information indicates that karst aquifers of
the Northern and Southern zones along the border area of Serbia and Bulgaria are still
not sufficiently studied, and cross-boundary groundwater flow is probably limited. A
number of obstacles exist to understanding of the cross-border relationships because
of limited access historical data due to the pre-existing border regime and the scarcity
of available information. The main difficulties are related to:
REFERENCES
Angelova D., Benderev A., Baltakov G., Ilieva I., Nenov T. (1995) On the Karst Evolution of
the Stara Planina Iskar Gorge. Review of the Bulgarian Geological Society 56(3), 111124
(in Bulgarian with English abstract).
Angelova D., Benderev A., Kostov K. (1999) On the age of the caves in the Stara Planina Iskar
gorge, NW Bulgaria. Proc. Karst 99 Europ. Conference, 1015. 09. 1999, Grand Causses
Verkors, France. 2935.
84 Karst without Boundaries
Awawdeh M., Nawafleh A. (2008) A GIS-based EPIK Model for Assessing Aquifer Vulnerability
in Irbid Governorate, North Jordan. Jordan Journal of Civil Engineering 2(3), 267278.
Antonov H. (1963) Karst waters in Western part of the Sofia Stara Planina. Annual University
of Mining and Geology, Sofia, 19611962, v.VII (in Bulgarian).
Antonov H., Danchev D. (1980) Ground waters in Bulgaria. Sofia, Tehnika, 360 p. (in
Bulgarian).
Directorate for Water Management of Danube region (2015) URL: http://www.bd-dunav.org/
content/Registar-na-izdadenite-razreshitelni-ot-podzemni-vodi-42/; last visited 28/4/2015.
Benderev A. (1989) Karst and karst waters of Ponor Mountain. PhD Thesis, NIPI-KG, 157 p.
(in Bulgarian).
Beron P., Daaliev T, Jalov Al (2006) Caves and Speleology in Bulgaria. PenSoft, Bulgarian
Federation of Speleology and National Museum of Natural History, Sofia. Bulgaria. 507 p.
CORINE Map. (2010) The European Topic Centre on Land Use and Spatial Information.
European Environment Agency. Retrieved from: http://www.eea.europa.eu/data-and-maps/
data/corine-land-cover-2006-raster-3.
Cvijic J. (1895) Caves and underground hidrography in Eastern Serbia. Glas. Serbian Royal
Academy of Sciences, XLVI, Belgrade, 1101 (in Serbian).
Calic J. (2007) Karst research in Serbia before the time of Jovan Cvijic. Acta Carsologica 36(2):
315319.
Doerfliger N., Jeannin P. Y., Zwahlen F. (1999) Water vulnerability assessment in karst environ-
ments: a new method of defining protection areas using a multi-attribute approach and GIS
tools (EPIK method), Environmental Geology 39 (2), 165176.
Dukic D. (1975) Characteristics of hydrology Serbia. Journal of the Geographical Institute
Jovan Cvijic, SANU, vol. 26, 2394 (in Serbian).
Eftimi R., Benderev Al (2007) Utilization of hydrochemical data for characterization of the
karst system & Example of Iskrets springs, Bulgaria. Review of the Bulgarian Geological
Society, part 13, 167174.
Gams I. (1994) Types of the poljes in Slovenia, their inundations and land use. Acta Carsologica,
XXIII: 285302.
Geological Survey of Serbia (1970) Basic geological map of Yugoslavia. Sheet Pirot. M
1:100000, Belgrade.
Gogu R. C., Dassargues A. (2000) Sensitivity analysis for the EPIK method of vulnerability
assessment in a small karstic aquifer, southern Belgium. Hydrogeology Journal (8), 337345.
Goldscheider N. (2005) Karst groundwater vulnerability mapping: application of a new
method in the Swabian Alb, Germany. Hydrogeology Journal 13(4), 555564.
Haidutov I., Dimitrova R. (editors) (1993) Geological map of Bulgaria. M 1:100000. Sheet
Berkovitsa, VTS, Troyan.
Haidutov I., (ed.) (1993) Geological map of Bulgaria. M 1:100000. Sheet Pirot, VTS, Troyan.
Hristova N. (2010) Hydrological picture of Niava trans-boundary catchment. Journal of the
Geographical Institute Jovan Cvijic SANU 60(2), 114.
Iurkiewicz A., Horoi V., Popa R. M., Dragusin V., Vlaicu M., Mocuta M. (2005) Groundwater
vulnerability assessment in a karstic area (Banat Mountains, Romania) Support for water man-
agement in protected areas, In: Stevanovic Z. and Milanovic P. (eds) Proceedings of International
Conference Water Resources and Environmental Problems in Karst CVIJIC, Beograd-Kotor,
p. 127133.
Koleva E., Peneva R. (1990) Climatic guide. The rainfalls in Bulgaria. BAS publishing house,
Sofia, 169 p. (in Bulgarian with English summary).
Kostov K. (2008) Indices of paleoseismicity in karst terrains. PhD thesis, Sofia, 191 p. (in
Bulgarian).
Mihaylova B., Kostov K., Danailova M., Benderev A. (2008) Characteristics and develop-
ment of the karst in the catchment area of the Opitsvet-Bezden springs. Proc. Conference
Development and protection of transboundary karst and karst aquifers 85
Investigation and protection of the Karst and Caves, Sofia, October 24, 2008, 96103 (in
Bulgarian with English abstract).
Mihaylova B., Benderev A., Kostov K. (2009) Application of GIS for localization of agricultural
land within the vulnerable areas of karst terrains (the example of Western Stara Planina).
GEOSCIENCES, Bulgarian Geological Society, Sofia. 125126 (in Bulgarian).
Mihaylova B., Benderev A. (2010) Characteristics of formation of karst water regime in the
southern slopes of the western part of Balkan Mountain (Bulgaria). In: Groundwater resources.
Current problems of learning and use. International Conference for the 100th anniversary of
the birth of B.I.Kudelin, Moscow, 1314 May 2010. 331337.
Paskalev M., Benderev A., Shanov S. (1992) Tectonic conditions of the region of the Iskrets
Karst Springs (West Stara-Planina). Review of the Bulgarian Geological Society 53(2), 6981
(in Bulgarian).
Petrovic J. (1974) Karst of eastern Serbia. Posebna izdanja Srpskog geografskog drustva,
Belgrade, 96 p. (in Serbian).
Petrovic J. (1976) Potholes and caves in Serbia. Vojnogeografski zavod, Belgrade, 511 p. (in
Serbian).
Radev J. (1915) Karst forms in Western Stara Planina. Ann. of Sofia University, Historical and
Philological Fac., v. 1011, 149 p. (in Bulgarian).
Ristic V. (2007) Development of simulation model for daily spring discharge estimation of karst
springs, PhD thesis, University of Belgrade, Faculty of Mining and Geology, Belgrade, 319 p.
(in Serbian).
Shanov S., Benderev A. (2005) Seismic impact from earthquakes and from grouped blasts
in quarries on the discharge of karst springs. In: Stevanovic Z. and Milanovic P. (eds)
Proceedings of International Conference Water Resources and Environmental Problems in
Karst CVIJIC, Beograd-Kotor, p. 145150.
Shanov S., Kostov K. (2015) Dynamic tectonics and karst. Springer-Verlag Berlin Heidelberg,
123 p.
Stevanovic Z. (1991) Hydrogeology of Carpathian-Balkan karst of Eastern Serbia and water
supply opportunities. University of Belgrade Faculty of Mining and Geology, Belgrade,
245 p.
Stevanovic Z. (1994) Karst ground waters of Carpatho-Balkanides in Eastern Serbia. In:
Stevanovic Z., Filipovic B. (eds.) Ground waters in carbonate rocks of the Carpathian-
Balkan mountain range. Spec ed. CBGA, Jersey-Belgrade, 203237.
Szeidovitz Gy., Paskaleva I., Gribovszki K., Kostov K., Surnyi G., Varga P., Nikolov G. (2008)
Estimation of an upper limit on prehistoric peak ground acceleration using the parameters of
intact speleothems in caves situated at the western part of Balkan Mountain Range, north-
west Bulgaria. Acta Geodaetica Geophisyca Hungarica 43(23), 249266.
Tchoumatchenco Pl., Rabrenovic D., Radulovic Vl., Maleevic N., Radulovic B. (2011) (a)
Trans-border (north-east Serbia/north-west Bulgaria) correlations of the Jurassic lithostrati-
graphic units. Geoloki anali Balkanskoga poluostrva 72, 120.
Tchoumatchenco Pl., Rabrenovic D., Radulovic Vl., Maleevic N., Radulovic B. (2011) (b)
Trans-border (east Serbia/west Bulgaria) correlation of the morpho-tectonic structures.
Geoloki anali Balkanskoga poluostrva 72, 2127.
Toula F. (1882) Grundlinien zur Geologie des Westl. Balkan. Denkschr der K. Ak. D. Wiss.
Wien, 44.
Yanev S. (ed.) (1992) Geological map of Bulgaria. M 1:100000. Sheet Sofia, VTS, Troyan.
Zagorchev I., Kostadinov V. (ed.) (1991) Geological map of Bulgaria. M 1:100000. Sheet
Vlasotice & Breznik, VTS, Troyan.
ivanovic V. (2011) Pollution vulnerability assessment of groundwater examples of karst in
Serbia, Master thesis, University of Belgrade, Faculty of Mining and Geology, Belgrade (in
Serbian). 215 p.
86 Karst without Boundaries
ivanovic V., Dragiic V. (2013) Application of EPIK method for vulnerability assessment of
groundwater of south part of Stara Planina Mt. Proceedings of 7th Symposium on Karst
Protection (in Serbian), J. Calic (ed), Beograd, p 1321.
Zlatkova M. (2006) Region 205 Bezden. Geological-geographical notes. Golyamata Temnota
Cave, village of Drenovo, Sofia district. Caves and Cavers, Sofia, 5556.
Zwahlen F. (ed) (2003) Vulnerability and risk mapping for the protection of carbonate (karst)
aquifers. Final Report COST Action 620, available at: http://capella.unine.ch/chyn/php/
publica_intro.php.
Chapter 7
An assessment of
territory participation
in transboundary karst aquifer
recharge: A case study from
the Skadar Lake catchment
area
Milan Radulovic 1, Goran Sekulic 2,
Momc ilo Blagojevic 3, Jelena Krstajic 4 & Entela Vako 5
1
University of Montenegro, Podgorica, Montenegro
2
University of Montenegro, Podgorica, Montenegro
3
Ministry of Agriculture and Rural Development, Podgorica, Montenegro
4
University of Belgrade, Belgrade, Serbia
5
Polytechnic University, Tirana, Albania
ABSTRACT
Equitable sharing of groundwater resources between countries is a major challenge, especially
in highly karstified terrains. In the karstic terrains of the External Dinarides water division takes
place mainly below the surface, so maps of the spatial distribution of groundwater recharge,
created by multi-parameter GIS methods, could be useful as one of the bases for water manage-
ment. Such maps can be used for assessing the percentage in which the territories of bordering
countries participate in recharging the shared aquifer. For the purpose of equitably sharing the
karst aquifer on the south western edge of Skadar Lake (shared by Montenegro and Albania),
the KARSTLOP method was applied. The method had previously been calibrated accord-
ing to the nearby catchments of four terrestrial karst springs. The Recharge map obtained
for the catchment area of sublacustrine springs allowed assessment according to which
291.7 million m3 of water infiltrates annually in Montenegrin territory, and 19.2 million m3 in
the Albanian part.
7.1 INTRODUCTION
There are many countries that share karst aquifers. Political, cultural, socio-economic
and other differences between bordering countries make groundwater management
one of the most challenging issues today. Conflicts over the use of groundwater some-
times arise, even between municipalities within a single country. An absence of coor-
dinated data sharing between municipalities and countries using the same aquifer can
lead to deterioration in qualitative and quantitative groundwater status. In addition,
disputes between countries over transboundary water resources often escalate, in some
cases into armed conflicts. Adequate hydrogeological research and sustainable water
management can help reduce the extent to which such instances arise. The grow-
ing problems of water scarcity and the unequal sharing of water resources between
88 Karst without Boundaries
countries have prompted a number of authors to deal with this issue (Bittinger, 1972;
Hayton & Utton, 1989; Yamada, 2004; Eckstein & Eckstein, 2005; Jarvis et al.,
2005; Puri & Aureli, 2005; Earle et al., 2010; Puri & Struckmeier, 2010; Ganoulis &
Fried, 2010; IGRAC, 2009; Stevanovic et al., 2012; Jarvis, 2014).
The hydrogeological research of transboundary aquifers in highly karstified ter-
rain is an especially complex task. The aim of this chapter is to present an approach
that will facilitate an assessment of groundwater resources in those areas. The External
Dinarides, to which the pilot site belongs, are characterised by holokarst (highly karst-
ified terrains), which demands a specific approach to hydrogeological research.
The approach presented is based on assessment of the spatial distribution of
groundwater recharge using GIS techniques. The focus of the approach is on the
application of the KARSTLOP method (Radulovic et al., 2012) as a tool for mapping
the spatial distribution of recharge in karst terrains. The approach is applied to the
transboundary catchment area of the sublacustrine springs of the south western coast
of Skadar Lake, which is shared by Montenegro and Albania. Previously, the method
was calibrated through analysis of the nearby catchment areas of terrestrial karst
springs (Podgor, Karuc, Slatina and Crnojevica springs) (Figure7.1).
7.2 METHODOLOGY
In highly karstified terrains, with numerous surface and subsurface karst landforms,
and with the absence of surface runoff, it is not possible to determine the spatial distri-
bution of aquifer recharge using standard methods (e.g. lysimeter, seepage meter, soil
mass balance, zero-flux plane, tracer methods, method based on the Darcys law, water
budget methods). The standard methods can only provide assessment of recharge
rates that relate to specific points or a whole catchment area of a karst spring.
Maps of the spatial distribution of groundwater recharge (recharge maps)
created by multi-parameter GIS methods (Shaban et al., 2005; Andreo et al., 2008;
N
D
IN
Montenegro AR
ID
ES
d
Bo c
ka b
ba
y a Sk
ad
AD ar
e lak
RI e
AT
IC
SE Albania
A
0 50 km
Figure7.1 Position of catchment areas: a) Podgor springs; b) Crnojevica springs; c) Karuc springs;
d) Slatina springs; e) the catchment area of the southwestern edge of Skadar Lake.
An assessment of territory participation in transboundary karst aquifer recharge 89
Plan Country
border
Divide
Qt = Ql + Qll
Country
Profile border Divide
Spring
Figure7.2 Conception of an assessment of the percentage in which the territories of bordering coun-
tries participate in the recharge of a shared karst aquifer; Q discharge (L3/s), A size of
catchment area (L2), Precipitation (m/year), R recharge (%).
90 Karst without Boundaries
In the process of obtaining the recharge map, the eight most important natural factors
affecting the recharge rate are analysed. The initials for the selected factors make up the
acronym in KARSTLOP method (where, K karstfication; A atmospheric conditions;
R runoff, S slope; T tectonics; L lithology; O overlying layers; P plants).
K map (Karstification) is created by analysing surface (Ksf) and subsurface karst-
ification (Kss). Surface karstification (Ksf) is assessed on the basis of the area of karst
landforms per surface unit (subfactors Ksf1 and Ksf2). When estimating subsurface
karstification (Kss), each side of a speleological object (swallow holes, caves, pits,
etc.) is marked in the range of 200 m, while the rest of the catchment area is assessed
on the basis of indirect parameters such as: the discharge amplitude of karst springs
(Kss1), the mean registered velocity of artificial tracers at catchment areas (Kss2) and the
primary mineral saturation index of spring water (Kss3).
A map (Atmospheric conditions) is obtained using the Digital Elevation Model
(DEM), which is first used to generate an Altitude map A1 and then a Shaded relief
map A2 showing factors affecting the intensity of solar radiation. An altitude (A1) indi-
rectly reflects the air temperature (and thus also evapotranspiration) and precipitation
(an increase in altitude leads to an increase in precipitation and more intense ground-
water recharge). The intensity of solar radiation (A2) also influences the amount of
evapotranspiration and recharge. Sides of mountains that are in shadow (surfaces
with relative reflection lower than 0.5) for the mean annual sun position (at noon
during the equinox) are characterised by lower intensity of solar radiation, i.e. by
lower evapotranspiration and higher recharge rate.
R map (Runoff) is obtained based on the distribution of permanent streams at the
catchment area, which may indicate the permeability of karstified terrains.
S map (Slope) is obtained using DEM. Terrains with smaller slopes are described
as being more suitable for groundwater recharge.
Two subfactors are considered for the preparation of T map (Tectonics): the den-
sity of faults T1 and dip of strata T2.
L map (Lithology) is obtained based on the following lithological data: the type of
carbonate rock (subfactor a), bedding (subfactor b) and mineral-petrographic impu-
rities (subfactor c).
O map (Overlying layers) is obtained based on data relating to the type and thick-
ness of geological cover and soil.
P map (Plants) is obtained based on data relating to terrain vegetation cover.
By overlapping these eight maps according to the established algorithm a recharge
map is obtained (Figure7.3, Table 7.1).
The method was applied to the catchment area of the southwestern edge of Skadar
Lake shared by Montenegro and Albania.
4 K map Karstification
(K = 15 points)
4 R map Runoff
(R = 15 points)
2 S map Slope
(S = 15 points)
1 T map Tectonics
(T = 15 points)
4 L map Lithology
(L = 0.674.80 points)
1 P map Plants
(P = 15 points)
of the catchment area is approximately 185 km2. Around 173 km2 (93.5%) belongs to
the territory of Montenegro, with the rest in the territory of Albania (12km2, i.e. 6.5%).
The area of Skadar Lake has a modified Mediterranean climate characterized
by hot and dry summers and cold winters. Mean annual precipitation in this area is
around 2500 mm, and the mean annual air temperature is around 10C.
Vegetation over the catchment area is represented by forest, especially at higher
altitudes. As elevation decreases, the area covered by forest gradually transforms into
one characterised by the low vegetation typical for karst terrain (scrubby vegetation).
A significant part of the catchment area is covered by pasture, but vegetation is gen-
erally absent from the coastal karst belt. The catchment area is characterised by the
presence of residual soil, the thickness of which is below 15 cm, with the exception of
areas with smaller slopes, where the thickness reaches 60 cm.
Geomorphologically, the area can be divided into two parts: the north eastern
slopes of Rumija Mountain and the Basin of Skadar Lake (Figure7.4). The north
eastern slopes of Rumija are made up of carbonate rock masses in which numerous
surface and subsurface karst landforms have developed. The most prevalent land-
forms are karrens that cover a significant portion of the surface. Sinkholes and uvalas
are also widespread, and the bottoms are generally slightly inclined in a strata dip
Table 7.1 Classification matrix for the KARSTLOP method (Radulovic et al., 2012).
K KARSTIFICATION
Area of degraded
zone (karren fields, Area of karst
ruine-like relief etc.) Ksf1 depressions per Ksf2 Ksf = (Ksf1 + Ksf2)/2
per unit square unitsquare
(103 m2/km2) (103 m2/km2)
Ksf + Kss
<60 1 <25 1 1
K=
60120 2 2550 2 >12 2
120180 3 5075 3 >23
180240 4 75100 4 >34
>240 5 >100 5 >45
Figure 7.4 The south western coast of Skadar Lake photographed from Montenegrin territory.
direction, i.e. to the north east. Also, a large number of caves can be found in this
area. Skadar Lake represents a crypto-depression whose floor in the areas of sublacus-
trine springs descends below sea level.
Many researchers both at home and abroad have studied the geological structure of
this region (Tietze, 1884; Baldacci, 1886; Cvijic, 1899; Nopcsa, 1916; Bourcart, 1926;
Waisse, 1948; Milovanovic, 1965; Beic, 1969; Grubic, 1975; Mirkovic et al., 1985).
The main base for exploring the geology of this area is a Geological map of
Montenegro 1:200 000 (Mirkovic et al., 1985). The Skadar Lake area belongs to the
tectonic unit of the Visoki Kr. Mesozoic rocks show dominant distribution and they
are represented mainly by carbonates, i.e. limestones and dolomites. The total thick-
ness of these carbonate rocks can be over 3000 m.
Numerous hydrogeological studies have been conducted in the Skadar Lake catch-
ment area (Torbarov & Radulovic, 1966; Radulovic, 1989, 2000, 2012; Radulovic
et al., 1979, 1989, 1998, 2013, 2015; Zogovic, 1992; Radulovic & Radulovic, 2004;
Stevanovic et al., 2008; Djordjevic et al., 2010; Sekulic & Bushati, 2013).
The surface drainage network in this area is poorly developed, with a small number
of streams that flow only in the rainy period of the year. In the study area the karst
aquifer is developed and it is mainly recharged by diffuse infiltration of rainwater
(autogenic recharge). The hydraulic conductivity of the carbonate aquifer is relatively
high. Groundwater flows mainly in privileged directions marked by faults and joints.
Discharge zones are spread along the south western edge of Skadar Lake.
Sublacustrine springs (vruljas), that occurring along the coastal part of the lake
(Figure7.5), are actually underwater dolines through which a carbonate aquifer
discharges. Around 40 sublacustrine springs are registered in the bottom of Skadar
94 Karst without Boundaries
6600000 6620000
4680000
4680000
N
Legend
Sublacustrine spring
4660000
4660000
Rain gauge
Divide
Isohyet
Lakeshore
Stream
Country border 0 5 km
6600000 6620000
Figure 7.5 Map of known sublacustrine springs on the south western edge of Skadar Lake with the
isohyets and the locations of the rain gauges.
Lake. The bottom of the underwater dolines is relatively deep, ranging from 10 to
70 m. One feature of karst aquifers in this area is a significant fluctuation in yield.
Some sublacustrine springs, such as the Radu and Krnjice springs, show a very high
yield in the rainy period (above 30 m3/s). Groundwater quality is relatively good, with
a high concentration of calcium and hydrocarbons.
Through statistical analysis of the digital recharge map (Figure7.6) the aver-
age recharge rate for the entire catchment area was obtained (68.5%). Using the
same map, in combination with other water balance components, the percentage of
territory participation in recharging the transboundary karst aquifer was assessed
(Table7.2). The results indicate that 93.8% of water that outflows through sublacus-
trine springs infiltrates via Montenegrian territory (291.7 million m3/year), and 6.2%
on that belonging to Albania (19.2 million m3/year).
7.4 DISCUSSION
The results obtained through applying the KARSTLOP method represent one of the
bases for further planning and sustainable management of the transboundary karst
aquifer shared between Montenegro and Albania.
4680000
Legend
Recharge (%)
4670000
4045
SK
AD 4550
AR 5055
LA 5560
KE
6065
6570
7075
7580
4660000
8085
0 5000 10000 m
Figure 7.6 Recharge map of the south west edge of Skadar Lake (after Radulovic et al., 2015).
Table 7.2 Balance components for the catchment area of the south western edge of Skadar Lake on
the territories of Montenegro and Albania.
The calibration of the method from the catchment areas of nearby terrestrial
springs proved applicable for assessing the spatial distribution of the mean recharge
rate. The calibration process shows that the error in assessing the recharge rate could
be up to around 5%. The discrepancy between calculated and previusly assessed
values was in range from 0.1 to 5.1% (Radulovic et al., 2012). However, the method
has been tested only on the four catchment areas, so that the possible error could
exceed the value.
However, as all methods of this type involve approximation, it is necessary to
have supporting data to reduce the subjectivity factor to a minimum. As the main lim-
itations of the KARSTLOP method, Radulovic et al. (2012) identified the following:
the method cannot assess allogenic recharge (inflows from non-karstic terrains
must be assessed through the application of other hydrological methods),
the method cannot assess temporal variations of the recharge rate (only the mean
annual value).
To apply the approach described it is important first to determine the spatial posi-
tion of the divides (watersheds), which is a difficult task in highly karstificated terrain.
However, if data related to the discharge of springs and the amount of precipitation
exists, the method can be used inversely, i.e. the recharge map can be used to find the
spatial position of divides.
In applying the method described it is not possible to assess static groundwater
reserves (storage), only dynamic reserves, i.e. the amount of water which inflows into the
system through the process of infiltration (recharge) and outflows via springs (discharge).
Based on results obtained through previous application, use of the method is
recommended for highly karstic areas, but the weighting coefficient of factors should
be adapted to the characteristics of a given exploration area, i.e. the method shouldbe
additionally calibrated (Radulovic et al., 2012).
The main advantages of this method are that the input data are readily available
and application is facilitated using GIS techniques.
7.5 CONCLUSIONS
The application of the approach has proven to be useful for the purpose of equitably
sharing water resources in highly karstified terrains. For applying the method in other
karst areas, it is necessary to bear in mind the details of the terrain and customise the
appropriate algorithm accordingly.
Use of the KARSTLOP method has generated one of the bases for the delinea-
tion and rational management of the transboundary karst aquifer shared between
Montenegro and Albania.
The approach could find a wider application in assessing participation in karst
aquifer recharge for other territories. Potential areas for the application of this
approach could be found in the Balkans, where the distribution of carbonate rocks
and the fact that a large number of countries exist within a relatively small area
caused the existence of transboundary karst aquifers.
An assessment of territory participation in transboundary karst aquifer recharge 97
Examples can be found in the cross-border areas of Croatia and Slovenia, Croatia
and Bosnia and Herzegovina, Bosnia and Herzegovina and Montenegro, Serbia and
Montenegro, Serbia and Bosnia and Herzegovina, and Montenegro and Albania. It is
necessary to analyse karst aquifer potential in order to avert conflicts over water use.
The application of the approach could help to improve water management among
these countries.
It always should be kept in mind that the method provides assessed recharge rates
on a regional scale, so results should be used with caution, and preferably compared
with values obtained by comprehensive hydrogeologic assessments.
Since water management for transboundary karst aquifers is a highly important
and sensitive issue, this topic should be considered carefully in future research, with
the participation of experts from various fields.
REFERENCES
Andreo B., Vas J., Durn J., Jimnez P., Lpez-Geta J.A., Carrasco F. (2008) Methodology for
groundwater recharge assessment in carbonate aquifers: application to pilot sites in southern
Spain. Hydrogeology Journal 16, 911925.
Baldacci L. (1886) Ricognizione geologicomineraria del Montenegro. Bollettino del Reale
Comitato geologico dItalia 17, 910.
Beic Z. (1969) Geology of MontenegroKarst of Montenegro (in Serbian). Geological Survey
of Montenegro, Podgorica.
Bittinger M.W. (1972) A survey of interstate and international aquifer problems. Ground Water
10(2), 4454.
Bourcart J. (1926) Nouvelles observations sur la structure des Dinarides adriatiques. In:
Actas del XIV Congreso Geolgico Internacional, Madrid, 18991941.
Cvijic J. (1899) Glacial and morphological studies of the mountains of Bosnia, Herzegovina
and Montenegro (in Serbian). Voice of the Serbian Royal Academy 57, 1196.
Djordjevic B., Sekulic G., Radulovic M., aranovic M., Jacimovic M. (editors) (2010) Water
potentials of Montenegro (in Serbian). Montenegrin Academy of Sciences and Arts, Podgorica.
Eckstein Y., Eckstein G.E. (2005) Transboundary aquifers: Conceptual models for development
of international law. Ground Water 43(5), 679690.
Earle A., Jgerskog A., jendal J. (editors) (2010) Transboundary water management: princi-
ples and practice. Earthscan, London.
Ganoulis J., Fried J. (2010) Transboundary water resources managment Needs for a
coordinated multidisciplinary approach. In: Earle A., Jgerskog A., jendal J. (editors)
Transboundary water resources managment: principles and practice. Earthscan, London,
925.
Grubic A. (1975) Tectonics of Yugoslavia. Acta Geologica 41, 365384.
Hayton R., Utton A.E. (1989) Transboundary ground waters: The Bellagio draft treaty. Natural
Resources Journal 29, 663722.
IJC (Institut Jaroslav Cerni) (2001) Water master plan of Montenegro. Government of
Montenegro, Podgorica.
IGRAC (2009) Transboundary aquifers of the world, 1:50 000 000. Special edition for the
5th World Water Forum, Istanbul.
Jarvis T., Giordano M., Puri S., Matsumoto K., Wolf A. (2005) International borders, ground
water flow, and hydroschizophrenia. Ground Water 43(5), 764770.
98 Karst without Boundaries
Jarvis T. (2014) Contesting hidden waters-confilct resolution for groundwater and aquifers.
Routledge, New York.
Milovanovic B. (1965) Epeirogenic and orogenic dynamics in the area of External Dinarides
and the problems of paleokarstification and geological evolution of holokarst (in Serbian).
J Geol Surv Serbia 4(5), 544.
Mirkovic M., ivaljevic M., okic V., Perovic Z., Kalezic M., Pajovic M. (1985) Geological
map of Montenegro 1:200 000. The Republic Self Managing Community of Interest for
Geological Exploration of Montenegro, Titograd.
Nopcsa F. (1916) Begleitwortzur geol. Karten von Nordalbanien, Rasien und Ostmontenegro.
Fldtani Kzlny 46, 712.
Puri S., Aureli A. (2005) Transboundary aquifers: a global programm to asses, evaluate, and
develop policy. Ground Water 43(5), 661668.
Puri S., Struckmeier W. (2010) Aquifer resources in a transboundary context: a hidden resource?
Enabling the practitioner to see it and bank it for good use. In: Earle A., Jgerskog A.,
jendal J. (editors) Transboundary water management: principles and practice. Earthscan,
London, 7390.
Radulovic V. (1989) Hydrogeology of Skadar Lake watershed (in Serbian). Geological Survey
of Montenegro.
Radulovic M (1994) Report on hydrogeological exploration of Sjenokos locality in Crmnica
area I phase (in Serbian). University of Montenegro, Podgorica.
Radulovic M. (2000) Karst hydrogeology of Montenegro (in Serbian). Geological Survey of
Montenegro, Podgorica.
Radulovic M.M. (2012) Multi-parameter analysis of groundwater recharge in karstic aquifers
case examples from the Skadar Lake basin (in Serbian). PhD thesis, University of Belgrade,
Belgrade
Radulovic M., Radulovic V. (2004) Hydrogeological map of Montenegro 1:200000 (in
Serbian). Geological Survey of Montenegro, Podgorica.
Radulovic M., Radulovic V., Popovic Z. (1979) Final report on the task of preparation for
the Basic hydrogeological map of SRY 1:100,000, Titograd sheet (in Serbian). Geological
Survey of Montenegro, Titograd.
Radulovic M., Popovic Z., Vujisic M., Novakovic D. (1989) Basic hydrogeological map of SRY
1:100,000, Bar sheet (in Serbian). Geological Survey of Montenegro, Titograd.
Radulovic M., Popovic Z., Vujisic M., Novakovic D. (1998) Guide book for the Basic hydro-
geological map of SRY 1:100,000, Bar sheet (in Serbian). Geological Survey of Montenegro,
Titograd.
Radulovic M.M., Stevanovic Z., Radulovic M. (2012) A new approach in assessing recharge
of highly karstified terrainsMontenegro case studies. Environmental Earth Sciences 65(8),
22212230.
Radulovic M.M., Novakovic D., Sekulic G. (2013) Geological and hydrogeological character-
istics of the Montenegrin part of the Skadar Lake catchment area (in Serbian). In: Sekulic
G, Bushati S (editors) Development of hydrogeological and hydraulic study of regulation of
Skadar Lake and Bojana river water regime (IPA Project)Volume 1. Montenegrin Academy
of Science and Arts, Podgorica, 9115.
Radulovic M.M., Radulovic M., Stevanovic Z., Sekulic G., Radulovic V., Buric M., Novakovic
D., Vako E., Blagojevic M., Devic N., Radojevic D. (2015) Hydrogeology of the Skadar
Lake basin (Southeast Dinarides) with an assessment of considerable subterranean inflow.
Environmental Earth Sciences 74(1), 7182.
Sekulic G., Bushati S. (editors) (2013) Development of hydrogeological and hydraulic study
of regulation of Skadar Lake and Bojana river water regime (IPA Project)Volume 1.
Montenegrin Academy of Science and Arts, Podgorica.
An assessment of territory participation in transboundary karst aquifer recharge 99
Shaban A., Khawlie M., Abdallah C. (2005) Use of remote sensing and GIS to determine recharge
potential zones: the case of Occidental Lebanon. Hydrogeology Journal 14, 433443.
Stevanovic Z., Radulovic M., Shammy P., Radulovic M.M. (2008) Karstic source Bolje Sestre:
Optimal solution for regional water supply of the Montenegro coastal area (in Serbian). In:
Zbornik radova Odbora za kras i speleologiju (Recueil des raports du Comite pour le karst
et la speleologie), Serbian Academy of Science and Arts, Belgrade, 9, 3364.
Stevanovic Z., Kukuric N., Treidel H., Peka Z., Jolovic B, Radojevic D., Pambuku A.
(2012) Characterization of transboundary aquifers in Dinaric karsta base study for sus-
tainable water management at regional and local scale. 39th International Association of
Hydrogeologists Congress, Niagara Falls, Canada, September 1621, 2012.
Tietze E. (1884) Geologische Uebersicht von Montenegro. Hlder, Vienna.
Torbarov K., Radulovic V. (1966) Regional hydrogeological research of Montenegro and
Eastern Herzegovina (in Serbian). Geological Survey of Montenegro, Titograd.
Waisse J.G. (1948) Les bauxites de lEurope centrale (Province dinarique et Hongrie). PhD
thesis, University of Lausanne, Lausanne.
Yamada C. (2004) Second report on shared natural resources: transboundary groundwater. UN
International Law Commission, UN Doc A/CN.4/539.
Zogovic D. (1992) Final report of hydrogeological exploration of Karuc source (in Serbian).
Energoprojekt, Beograd.
ivaljevic R. (1991) Hydrological analysis of karst water flowCrnojevica River case study (in
Serbian). PhD thesis, University of Montenegro.
ivaljevic R., Bokovic M. (1984) The results from hydrological measurements in the catchment
area of Orahovstica (in Serbian). Hydrological and Meteorological Service of Montenegro,
Podgorica.
This page intentionally left blank
Part 2
Karst aquifer
characterisation and
monitoring
This page intentionally left blank
Chapter 8
ABSTRACT
In the south-eastern Dinarides, with the highest precipitation in Europe, people still suffer
from water shortages. The Dinaric Karst Aquifer System (DIKTAS) belongs to one of the
largest karst areas extending along the coast of the Adriatic Sea from Slovenia to Albania.
Thelarge water potential of the DIKTAS area is mostly concentrated on the catchment
areas of the Cetina, Neretva, Trebinjica, Zeta and Bojana rivers. This potential underpins
an advanced economic development programme for the region was recognized long ago.
However, the karst phenomena always presented a significant barrier for groundwater use.
The karst presents a variety of hazards and risks associated with human activities, particularly
the construction of dams, reservoirs, tunnels and canals. Successful solutions require seri-
ous and complex approaches and close co-operation with a wide spectrum of scientists and
engineers to define causes and consequences between human activities and impact. The area
of south-eastern Dinarides, particularly Eastern Herzegovina, represents the optimal strategy
for water resources management in karst areas which is a key requirement for regional socio-
economic development.
8.1 INTRODUCTION
The area of south-eastern Dinarides is one of the most karstified regions in the world.
It has an average annual precipitation of 12502450mm, locally more than 5000mm,
and maximum of 8000mm (Figure8.1).
Sinking rivers, underground flows, temporary flooded karst poljes and lack of
arable land are the main properties of the region. Arable land, karst poljes, uvalas and
sinkholes are surrounded by bare rocks (Figure8.2).
The people of this region have always had to cope with two kinds of misfortune:
flood and drought (Figure8.3).
Under natural conditions a large portion of the arable land is flooded 150 to
250 days each year. Sowing and harvesting are not determined by man, but by water.
During dry summer periods villagers rely on rainwater collected during winter or on
water from siphon lakes in deep natural karst shafts and caves. For centuries people
have emigrated from this region in search of a better life. The population density per
km2 varies from 21 to 29. The only natural resource of importance for the regional
socio-economic development is water. However, distribution of precipitation is
104 Karst without Boundaries
uneven during the year. More than 70% occurs during the wet season (October
March). The aquifer retardation capacity is extremely poor. Ninety percent of
groundwater turnover occurs annually. The groundwater potential of all springs in
the area of Eastern Herzegovina and Boka Kotorska Bay is estimated with a yearly
Optimal water management 105
Figure8.3 Fatnica Polje under water (upper). Dry Trebinjica river bed (lower). Photos P. Milanovic.
average of 300 m3/s. No doubt, this region is the largest and most important fresh
water source in the Mediterranean area. However, only a small amount of this water
is successfully tapped.
More than 90% of the entire area consists of soluble Mesozoic carbonate formations.
The depth of these rocks exceeds 3000 m. The Eocene flysch does not represent a
significant lithostratigraphic unit, however, according to its hydrogeological role and
location along reverse faults, it has a huge effect on karst aquifers at many karst
poljes. The existing geo-structural features of the entire Dinaric Karst Aquifer System
(DIKTAS) area are the result of subduction movement by the Adriatic microplate
beneath the Dinarides mountain chain. The dominating tectonic stress is oriented
from the south west to north east and subsequently the regional structures are ori-
entated north west to south east. The main structural features are the thrust faults
locally separating the flysch from the carbonate complex limestones and dolomites.
Inclination of the reverse fault planes, dipping toward the north east, ranges between
55 and 75. Strong tectonic movements cause folding and cascade lifting of the ter-
rain parallel to the present-day coastline. Starting at an altitude of over 1000 m, these
106 Karst without Boundaries
stepwise depressions, which erosion and sedimentation have turned into spacious
karst valleys known as poljes, are the outstanding feature of the Dinaric, and of the
entire DIKTAS karst region.
Organisation of the water regime of this part of south east Dinarides is the only
way forward for regional economic development. To minimise floods and to keep
water at the surface as long as possible is a basic requirement for economic progress.
Protection of ponors (swallow holes) against natural plugging, construction of small
surface water storage ponds and construction of water-driven mills at ponors have
been ongoing for centuries. The aqueduct constructed in Roman times for the water
supply of Epidaurus (Cavtat) and the tapping of three karst springs for Dubrovnik
water supply are some of the ancient examples. The first dam in the DIKTAS area
(Klinje Dam, in Herzegovina) was constructed in 1896 for the Gatacko polje irrigation
system (Figure8.4). By excavation of many dewatering tunnels, after World War II,
a number of karst poljes in the DIKTAS region have been protected against floods.
Intensive construction of dams and reservoirs, mostly in Europe and the United
States, started in the first half of the 20th century. One of the first successful proj-
ects was the American Tennessee Valley Project. In that time understanding of karst
and karstification, from an engineering view point, was in its infancy. A number of
large structure failures were registered in this period. Dried reservoirs or reservoirs
with unacceptable heavy leakage were frequent (e.g. Hales Bar, US; Montejaque,
Spain; Vrtac, Montenegro). Failure in karst occurred despite extensive investigation
Large and successful reservoirs constructed in the Dinaric karst area after World War
II (Peruca at Cetina River, Bileca at Trebinjica River and Krupac + Slano at Zeta
River) promoted the Yugoslav school of engineering karstology. The Grancarevo
Dam and Bileca Reservoir, as part of the Trebinjica Multipurpose Hydrosystem
(TMH), are excellent examples of successful construction of large structures in
hazardous a and unpredictable karst geological environment (Figure8.5). This
Hydrosystem is situated in the Eastern Herzegovina, in the middle of the DIKTAS
project area.
Stepwise placement of karst poljes in Eastern Herzegovina allows optimal mul-
tipurpose use of the great water potential from an elevation 1000 m down to the sea
level. The water storage at the surface from the rainy period of year allows continued
use during the drier months. This goal can be achieved only by construction of dams
and reservoirs interconnected by tunnels and canals, and the construction of an inte-
grated regional hydrosystem. The Trebinjica Multipurpose Hydrosystem was initi-
ated in early 1950s. The idea is to construct a huge surface water way in the shape
of letter S (Milanovic, 2010) (Figure8.6).
In this way part of the water potential will be accessible to the majority of the
population in the region, for different purposes, through the year. The impacts from
0 2 4 6 8 10 20 km
1
2
3
4
A
5 d
r
ia
6
t
ic
7
s
8 11 e
a
9 12
10 13
floods and droughts will be considerably reduced. The first governmental document
for the construction of this integral reclamation and hydropower system was accepted
in 1960. All structures of the Hydrosystem have been designed in such a manner that
the water potential could be used for food production; water supply; flood reduction;
irrigation; hydropower production; industry; fish farming; recreation; and prevention
Optimal water management 109
of deforestation. It was also designed to have a positive influence on the critical min-
imal flows in urban and nature protected areas with a number of secondary benefits,
including the most important benefit to decrease the strong emigration trend away
from the region.
The TMH (Figures 8.7 and 8.8) consists of seven dams, six reservoirs, six tunnels
(with a total length of 74 km) and four canals (with a total length of 74 km). A large
part of TMH (Phase I) is already operational.
The Phase I includes waters from the Trebinjica springs which are at sea level.
The Trebinjica River is the most important in the area as it is the longest sinking river
in Europe with a total length of 90 km of which about 30 km has permanent flow.
The TMH aims to harness the potential energy of this river. The main structures of the
Phase I are: Grancarevo Dam (123 m high); Bileca Reservoir (V = 1280 hm3); Gorica
Reservoir (V = 15 106 m3); Gorica Dam (33.5 m high); tunnel Gorica Dubrovnik PP
Legend
250 1 A
d
2 r
i a
3 t i
c
4 S
5 e
a
6
7 0 5 10 15 20 km
Figure8.7 General scheme of the Trebinjica Multipurpose Hydrosystem, layout. Altitude; 2. Tempo
rary flow; 3. Permanent flow; 4. Tunnel route; 5. dam; 6. power plant (operational); 7. power
plant (designed).
Altitude
PP NEVESINJE 60MW
GATACKO POLJE
1300
RILJA DAM
longitudinal profile
POSCENJE DAM
1200
1100
reservoir Zalomka
973 970 973
1000
PP DABAR 160 MW
D-4,6 m
GRANCAREVO DAM
900 L - 4,3 km 980
PP BILECA 33 MW
Dabarsko polje
791
PP TREBINJE l 180 MW
D-4,6 m
PP TREBINJE ll 8 MW
Fatnicko polje
800 L - 11,08 km
reservoir Nevesinje
700
PP DUBROVNIK 216 MW
600
476 470 D-5,0m 450
500 D-5,4 m reservoir Bileca
L - 3,24 km 421 420
L - 15,84 km
400
295
300 231.5 D-6,0 m 246
lower and upper compens. reservoir L - 16,6 km
200 CANAL L-68
D - 8,0 m
Reservoir Trebisnjica
100 L - 8,09 km
3 00
ADRIATIC Sea river NERETVA SEA
0
0 5 10 15 20 25 30 35 40 45 50 55 70 75 80 85 90 95 100 105 110 115 120
48
RPP CAPLJINA 420 MW L (km)
(16.57 km); Trebinjica river bed, paved by shotcrete (65 km); Hutovo Reservoir (V =
5 106 m3); tunnel for reversible PP Capljina (8.093 km), and power plants Trebinje I,
Trebinje II, Dubrovnik and Capljina.
The Phase II (so-called Upper Horizons) is now under construction. This part of
the project requires the transfer of 30% of the water from the upstream part of the
Neretva River (springs Buna + Bunica + Bregava, Qav = 59.6 m3/s), to be used at the
existing operational power plants (Phase I), and is then returned to the downstream
part of Neretva River. As part of the water re-routing three temporary flooded karst
poljes will be flood free and available for agriculture. According to the l hydrogeolog-
ical and hydrological analysis, including four mathematical models, it was confirmed
that the consequences of water transfer would be negligible on the Buna Spring dis-
charge. In the case of the Bunica Spring the maximum flow (high precipitation period
of year) would be reduced to about 70%, however in the period of low flow (dry
period of year) a negative influence of water transfer is not possible. A similar result
occurs in the Bregava Springs and Bregava sinking river. According to the design,
the geotechnical measures are intended to improve the watertightness of the pervi-
ous Bregava river bed, with the consequence that the flow in the dry period would
increase 3 to 5 times over the natural conditions (Milanovic, 2006).
Once completed the TMH will enable irrigation of about 240 km2 of arable land
and an average annual hydropower output of 856 GWh. The Bileca Reservoir, sit-
uated in karstic carbonates (volume 1.3 109 m3), without leakage, is one of largest
and most successful reservoirs in karst. The question of reservoir watertightness was
the key question of the entire project feasibility.
After construction the first phase of the TMH, the surface and underground water
regime at one of world famous karst phenomena Popovo polje has been drastically
changed (Figure8.9). The possibility of massive floods (up to 40 m deep), which fre-
quently happened under natural conditions, are now possible only at the very end of
polje, but the flood water is shallow and of short duration.
Figure8.9 Popovo polje before and after construction of first phase of the Trebinjica multipurpose
hydrosystem.
112 Karst without Boundaries
From the beginning the project was strongly supported by people from all communi-
ties from this part of the Dinarides and approved by responsible governmental insti-
tutions. At the same time the important question of possible environmental impact
arises as a consequence of natural water regime changes. Some large springs, tapped
for water supply (Trebinjica Spring and Oko Spring) were submerged by reservoirs
(Figure8.10). The possible impact of Trebinjica Spring submergence on reservoir
integrity was also one of important questions to be answered in the stage of design.
Concern over the possible negative impact on the flow regime downstream from
springs often creates conflicts between the owners of the reservoirs and the user of the
springs. This too needs addressing.
Karst is rich with various fauna including a number of endemic species. Changes
in the underground and surface water regimes can cause a negative effect on the fauna
of subterranean karst. The important issue was how to keep the balance between
the necessity for regional development and preservation of the sensitive and complex
karst ecological system (Milanovic, 2002).
Subterranean terrestrial and aquatic species exist in karst systems all over the
world. As a consequence of the natural underground and surface water regime distur-
bance, a very distinct impact on a wide spectrum of biodiversities is expected. Often,
construction of any structure in karst, particularly dewatering of temporarily flooded
karst depressions and construction of dams and reservoirs has resulted in negative
influence on fauna in the caverns.
In the East Herzegovina karst area a number of aquatic and terrestrial species have
been found. Particularly well known is Vjetrenica cave (Popovo Polje) with several
endemic species. A well-known cave-dwelling aquatic species is the blind salamander
Figure8.10 T
rebinjica Spring, dry and submerged by a 75 m depth of water, in period of high
discharge (Photos P. Milanovic).
Optimal water management 113
Proteus anguinus known as human fish is common in the Dinaric karst. At Popovo
Polje Proteus was found at more than 40 localities (Cuckovic, 1978). A problem arises
in the case of the fish Paraphoxinus ghetaldi that inhabit siphons of estavelles in the
Popovo Polje (Bosnia Herzegovina) during dry periods of the year and intermittent
lakes of the karst poljes during the flood season. After the karst polje is permanently
protected from flooding, fishing at the openings of estavelles, for centuries an import-
ant traditional food source for the local inhabitants, is no longer practical.
The unique tubeworm Mariphugia cavatica (Figure8.11) that colonises the tem-
porary flooded karst channels are seriously endangered after dewatering temporarily
flooded karst poljes. A large colony of this warm was discovered, also in a karst
channel beneath, 10 m of alluvium (Popovo Polje, Figure8.9). After construction of
the impervious blanket over the alluvial deposits to prevent seepage from the reservoir
bottom the survival of Mariphugia is questionable.
To avoid possible conflicts and to provide the sustainable and peaceful utilisa-
tion of transboundary waters the designers of TMH undertook a series of long last-
ing investigations. These were carried out collaboratively with a wide spectrum of
scientists and engineers including geologists, hydrologists, chemists, civil engineers, biol-
ogists, archeologists and seismologists. An inventory of 120 springs and a monitoring
programme began 8 years before the project was operational; a programme of regional
water quality monitoring has been organised; the possible influence of TMH on a shell
farm at the sea coast was investigated; while some important historical monuments
were displaced by the reservoir water (Figure8.12); submerged water sources were
replaced by new capturing structures; and number of different environmental analyses
were performed. In most instances the impact is positive and predictable. However, some
impacts are negative and unpredictable, such as, for example, karst induced seismicity.
With the initiation of the construction of large structures the technology of investi-
gating karst phenomena, including many new methods in hydrogeology, hydrology,
20 mm
curtains, cut-off walls, concrete plugs and underground dams. A number of these
methods has been applied during the construction of the different TMH structures.
During past 50 years the constructed part of the TMH has considerably improved
the quality of life.
One of the important lessons learned is that in karst engineering a certain degree
of flexibility is necessary during the construction of large structures in order to find
the best compromise between different possibilities and solutions, including protec-
tion of environment and historical heritage.
Experience and knowledge are important bases for further utilisation of the
groundwater potential, as water is presently lost to the sea through a number of per-
manent, temporary and submarine springs. Successful safe reservoirs and impervious
canals, as well as long tunnels, confirm high level of knowledge and high quality of
experience in engineering karstology.
The large water potential of this region will be of great importance in the future. The
water potential of the Eastern Herzegovina and Boka Kotorska Bay area is not only
of local importance but is internationally significant as well. This region would have
a role of the Mediterranean Water Treasure (Milanovic, 2004). New technologies
make possible tapping and control of groundwater flows deep underground.
By abstracting underground water in periods of high precipitation, at elevation
between 200 and 300 m, distribution is possible by gravity over large distances.
Particularly important is the large, but temporary stored amount of water in the Orjen
Mountain (Figure8.14).
Successful abstraction of only 20% of Orjen water provides 1015 m3/s through-
out the year. The Water Treasure concept requires permanent and perfect protection
of surface and underground water over the entire region of Eastern Herzegovina and
the wide area of the Boka Kotorska Bay. Eventual problems caused by transbound-
ary aquifers have to be overcome and the entire region should be treated as a unique
hydrogeological and hydrological entity.
8.8 CONCLUSIONS
Large sinking rivers, numerous sinkholes, ponors, temporarily flooded karst poljes
and deeply concentrated underground flows characterise the Eastern Herzegovina
karst region. In spite of abundant rainfall, karstic terrain and uneven distribution of
precipitation makes its inhabitants vulnerable to frequent floods and droughts. To
provide optimum water management the TMH was initiated as the best feasible solu-
tion. The ultimate aim was to improve the livelihoods of the people in the region by
regulating water availability all year round for all users. The Reservoir Bileca was the
first reservoir with a volume of more than 1 billion cubic meters of water constricted
entirely over heavily karstified rock. There was no seepage. The TMH is ongoing
and is the most important agent of economic development in Eastern Herzegovina.
116 Karst without Boundaries
Map of izohyet
Average annual Precipitation in the area of Orjen Mauntain
Ad
ri
at
ic
Se
a
Precipitation station
Large karstic spring
0 Precipitation in mm 0 10 20 km
250
However, the environmental properties of complex karstic system and the state bor-
ders dissecting the region make integrated management of water resources extremely
complex. The need for optimal use of the large water resources still calls for a great
effort by all users in the region and for a holistic approach.
REFERENCES
Cuckovic S. (1978) The question of the survival of well known endemic Proteus Anguinus in
the Trebinjica Power System. Conference on environmental effects of storage reservoirs.
Yugoslav Committee for Large Dams. Trebinje.
Milanovic P. (2000) Geological engineering in karst. Zebra, Publishing Ltd, Beograd.
Milanovic P. (2002) The environmental impacts of human activities and engineering construc-
tions in karst regions. Episodes, 25(1), 1327.
Milanovic P. (2004) Water without boundary water resources potential in deep karst of South-
Eastern Dinarides. Book of abstracts 32th International Geological Congress, Florence, Italy.
Milanovic P. (2006) Karst of Eastern Herzegovina and Dubrovnik littoral. ASOS, Belgrade.
Milanovic P. (2010) Transboundary aquifers in karst source of water management and political
problems. Case study, S-E Dinarides. ISRAM 2010 International Conference. Transboundary
aquifers. UNESCO, Paris.
Chapter 9
ABSTRACT
Spring hydrograph recession analysis is a well-known topic in karst hydrogeology. The different
models in the literature can be separated into empirical and physically-based models; in the
latter, only analytical models provide the discharge equation during recession. Under constant
geometrical and hydraulic aquifer characteristics, the exponential form appears to be the most
recurrent theoretical type, at least during the long-term flow recession. During this stage, any
deviation from the exponential form, may suggest hydraulic anisotropy of the aquifers, as well
as aquifer geometry which has a fundamental role in controlling the shape of spring hydro-
graphs. Karst aquifer hydraulics, under different hydrological conditions, are described along
with the relationship to different spring hydrograph segments.
9.1 INTRODUCTION
Spring discharge hydrographs can be used to define the regime of springs, and shape
provides a useful tool to investigate some of the karst features of the aquifers. Well
developed karstification results in spring hydrographs characterised by pronounced
peaks, which follow the main rainfall events, and indicate both a good connection of
the conduit networks with the spring and the presence of active recharge pathways.
A smoother shape, without sharp peaks and characterised by one or few flood peaks
during the hydrological year, indicates a coarse development of karst conduits or that
they are poorly connected.
During the recession, the hydrograph is generally the most stable, and it is believed
to express some geometrical and hydraulic characteristics of aquifers. This has stim-
ulated numerous studies, and hydrograph recession analysis is a well-known topic in
hydrogeology. There are a large number of studies recorded in the literature, and these
try to describe the analytical expressions of the recession and their physical meaning
(Dewandel et al., 2003; Bailly-Comte et al., 2010; Fiorillo, 2014).
Figure9.1 is a theoretical example of the spring hydrograph and its recession limb.
In this example, an intense pulse is transmitted through the system, which causes the
water to rise in spring. The recession limb is considered here as an exponential function,
characterised by the response time, tR, which is the inverse value of the recession coef-
ficient, (tR = 1/). The beginning of the recession would occur at the inflection point,
Q0, which splits a convex (left side) and concave (right side) form of the hydrograph.
Hydrograph recession can occur for short periods (several days or weeks) between two
main rainfall events, but has a longer duration in the dry season.
118 Karst without Boundaries
Rising limb
Recession limb
tL Q0 where d2Q/dt2 = 0
Q=Q0 e t
Discharge, Q
=1/tR
tB
Base flow = QB
Time, t
Figure9.1 Different parts of a spring hydrograph triggered by a storm pulse (modified from White,
1988): tL, length of time between the storm pulse and the peak in the discharge hydrograph;
tB, time to return to base flow; a, recession coefficient; tR, response time.
Q(t) = Q0 - T t (9.1)
Q(t) = Q0 e - t (9.2)
where Q0 is the initial discharge at time t = 0, and the recession coefficient T and depend
only on the aquifer hydraulic systems; they are constant during the recession period.
Equation 9.1 describes the linear decrease of discharge with time; the main conse-
quence of this behaviour is the constant slope of the discharge-time plot at any time,
independently of the initial value, Q0. This equation is obtained by draining a cylin-
drical tank-reservoir, where no energy is lost during the emptying (Table9.1).
Equation 9.2 was first provided by Maillet (1905) to describe the spring drainage
from a porous aquifer, obtained from a cylindrical tank with a porous plug. The phys-
ical meaning of the recession coefficient, , using different hydraulic laws to drain
Spring hydrograph recession: A brief review focused on karst aquifers 119
4000 120
hydrological year
wet season dry season
3500 100
intra-annual recession
daily rainfall, mm
discharge, l/s
80
3000 seasonal recession
60
2500
40
2000
20
1500 0
1-Sep
1-Dec
1-Mar
1-Jun
1-Sep
1-Dec
Figure9.2 Example of karst spring hydrograph recession in a Mediterranean climate: Torano daily
spring discharge (Matese Mountain, southern Italy) and daily rainfall recorded in the
recharge area (Lago Matese rain gauge, annual mean 1948 mm). Intra-annual recession
occurs during the wet season, interrupted by the next rainfall event; during these peri-
ods, recharge processes progressively increase the basal water table level of the aquifer.
A seasonal recession begins when recharge processes decrease as a consequence of tem-
perature increase and rainfall decrease throughout the (Mediterranean) summer and can
extend in the following hydrological year (modified from Fiorillo, 2014).
Table9.1 Discharge equations found from draining a cylindrical tank-reservoir using different physical
laws (from Fiorillo, 2011). Application of Darcys Law is carried out filling the connected
tank-tube (with area A2 and length L) by sand with hydraulic conductivity K. Application of
Poiseuilles Law is carried out considering a small diameter d of the connected tank-tube;
is the dynamic viscosity of water; is the water density; g is the gravity acceleration.
Tank shape Physical law Discharge equation Recession coeff. Discharge for t = 0
A22
Torricelli Q( t ) = Q0 - T t T = g Q0 = A2 2 gh0
A1
A1
h0 K A2 KA2 h0
Darcy Q( t ) = Q0 e - D t D = Q0 =
L A1 L
L A2 d4 g d 4 g h0
Poiseuille Q( t ) = Q0e - P t P = n Q0 = n
128 L A1 128 L
As it gives a straight-line in the semilogarithm plot, the exponential form of the reces-
sion has found wide use in hydrogeology. It can describe a large part of the recession.
120 Karst without Boundaries
Real examples of the recession limb of karst spring hydrographs in the semilog-
arithmic plot have shown a variation in the angle of the straight line. This indicates
the presence of more than one recession coefficient during the entire recession period
(Forkasiewicz & Paloc, 1967; Milanovic, 1976; Atkinson, 1977).
Other equations describing the recession exist and differ from the linear and
exponential form.
The main models analysing the spring hydrograph recession have been discussed in
Fiorillo (2014). They vary from pure empirical to physically-based models. In the
last ones, only analytical models provide the discharge equation during the recession
phase. Most models which analyse the spring hydrograph recession can be considered
as lumped (or global) models, as the flow of the total catchment is often focused on
one single spring and, therefore, allows for an integral characterisation of the flow
behaviour of the whole catchment just by measuring the outflow at a single point
(Jeannin & Sauter, 1998). Only a few models are able to consider the spatial hetero-
geneity of the aquifer.
Most of theoretical models consider the hydrograph recession that is not condi-
tioned by recharge processes. Under this assumption, the only aquifer zone believed
to be involved is the saturated zone, and the role of the vadose zone is generally
neglected. However, during the recession, some recharge will reach the saturated zone
and influences the discharge of springs; this process occurs especially during the first
part of the recession and characterises the influenced-stage. Only the last part of the
recession, which is the most stable part of the hydrograph, could be considered as a
non-influenced stage and it is known as the recession limb.
Empirical models do not consider any geometrical and hydraulic characteristics
of the aquifers, as they are generally based on the best fits of experimental reces-
sion data; thus, they could not provide any information on the structure of aquifers.
However, many empirical models start from very simple reservoirs to obtain discharge
equations or try to obtain some hydraulic parameters of aquifers from hydrographs.
As these last models evaluate some physical characteristics of the aquifers, they can be
considered as semi-empirical models.
Physically-based models require a definition of the geometrical and hydraulic
conditions of the aquifers, and also need specific physical laws which describe the
drainage. These characteristics are very difficult or impossible to obtain for actual
karst aquifers, and for this reason many physical models are based on simple theoret-
ical geometric and hydraulic characteristics of the aquifers, which allow the equation
of the discharge to be obtained from analytical models. Most models were developed
for porous media, but they have been largely applied to karst media; in any cases
no-recharge is assumed in the models.
Forkasiewicz & Paloc (1967) characterised the karst aquifer with different paral-
lel linear reservoirs, all contributing to a spring hydrograph. They reconstructed three
segments of the hydrograph in a semilogarithm plot as the different contributions of
each reservoir: the conduit network, the intermediate system of well integrated karsti-
fied fissures, and the low-permeability network of pores and narrow fissures.
Spring hydrograph recession: A brief review focused on karst aquifers 121
Nutbrown & Downing (1976) stated that different exponential terms can derive
exclusively from groundwater flow dynamics. Szilagyi (1999) has highlighted that the
changing slope in the recession plot can be simply the consequence of baseflow drain-
age. Estrela & Sahuquillo (1997) also modelled the recession hydrograph by several
decreasing exponentials, without identifying different flow regimes. Eisenlohr et al.
(1997) showed that the sum of three exponentials can be fitted on the hydrograph of
a system consisting of only two classes of hydraulic conductivities.
Baedke and Frothe (2001) state that the largest slope of the hydrograph is related
to the conduit flow system and the smallest slope to the diffuse flow system (and
deriving the aquifer diffusivities from each slope following the approach proposed by
Rorabaugh, (1964) and Sahuquillo & Gmez-Hernndez (2003). They highlighted
that the different slopes are never representative of the response of a more or less
pervious portion of the aquifer.
Similarly, Kovacs & Perrochet (2008) showed that different exponential compo-
nents do not correspond to aquifer volume with different hydraulic conductivities as
they can be extracted from a 2D analytical solution for diffuse flux from symmetric
and asymmetric rectangular blocks.
These exampes indicate that it is possible to explain the different slope of the
hydrograph in the semilogarithmic plot without requiring different flow regime (in the
conduits, minor fracture, etc.). The exact analytical solution provides a shape of the
hydrograph which is a sum of different exponentials, as in the model of Rorabaugh
(1964), Brutsaert (1994) or Kovacs et al. (2005). Only the latest part of the hydro-
graph can be approximated to an exponential, as it follows an initial non linear part
on the semilogarithmic plot.
Most physically-based models are based on the assumption that the aquifer is
characterised by homogeneous and isotropic media and the formation constants
Physically-based models generally assume that flow is Darcian.
Table9.2 shows analytical solutions found draining two different tank-shapes:
the exponential function (Equation 9.2) can be obtained using the Torricelli Law, and
the linear function (Equation 9.1) can be obtained using the Darcy/Poiseuille Laws,
contrary to the cylindrical tank-reservoir of Table9.1. Table9.2 show that it is possi-
ble to back-calculate the variation of the shape of the reservoir to obtain the desired
hydrograph, but it is more difficult to fix the type of flow and the geometrical model
for an actual karst aquifer. Analysing an actual spring hydrograph is difficult when
Darcian flow occurs, and unclear how the aquifer geometry controls the recession.
Fiorillo (2011) noted that many of the equations for the recession coefficient
could be linked to:
c
= (9.4)
neff Ac
where neff is the effective porosity computed along the water table with an area Ac
and c is the hydraulic constant, involving the geometric and hydraulic characteristics
of the discharge zone which does not vary during the drainage. This means that,
when the exponential form is expected (during the long-term flow recession), any
variation of the recession coefficient would depend on the product neff Ac (Fiorillo,
2014).
122 Karst without Boundaries
Table9.2
Discharge equations obtained while draining tank-reservoirs that are characterised by
different geometry using different physical laws (extracted from Fiorillo, 2012): l, length of
the reservoir in the orthogonal direction of x; B, constant of the reservoir; C=A1/l for x=0;
other symbols are described in the text. Hyperbolic tank shape has a rectangular base area
A1=x l; triangular tank shape has a rectangular base area A1 = (C x ) l.
Discharge
Tank shape Physical law equation Recession coefficient
A1 = x l
A2 2g
Torricelli Q( t ) = Q0 e - t T =
2l B
B
h=
h0 x2
L
X A2
A1 = (C X) l 2
A K B
Darcy/ Q( t ) = Q0 - t D = 2
L l
arc
tan
(B)
Poiseuille
2
n d4 g B
P =
128 L l
h0
h(t) = (C X)B
L
A2
X
In(Q)
Q, m3/s
0.50
3 constant
1
2 1.00
Q 1
0.5 In(Q) 3
1.50
0 2.00
0 60 120 180 240 300 360
t, days
Figure9.3 Arithmetic and logarithmic plots of constant (curve 1) and non-constant (curve 2 and 3)
recession coefficient value a during emptying (modified from Fiorillo et al., 2012).
decay of discharge through time provides information on the actual emptying rate
of the aquifer compared to a simple exponential decline (Figure9.3). If the reces-
sion coefficient, , decreases under drought conditions (curve 2), springs guaran-
tee water during a long dry period and can be considered drought-resistant (Fiorillo
etal., 2012). This hydraulic behaviour could be connected to an increase of the effec-
tive porosity with depth. If the recession coefficient increases under drought con-
ditions (curve 3), a different hydraulic behaviour occurs, as the aquifer is drained
more quickly than expected, and the springs can be considered as drought-vulnerable
(Fiorillo et al., 2012). This hydraulic behaviour could be connected to a decrease of
the water table area, Ac, or to the fact that the aquifer is also drained by other springs
with lower ground-elevation.
The different hydraulic behaviour of springs suggests that the common and dif-
fuse computation of the water stored in a karst aquifer obtained by integrating the
Maillet formula (Equation 9.2), over the time interval t0 to t, should be used with
caution in the evaluation of water resources, as it underestimates or overestimates the
volume discharged for drought-resistant and drought-vulnerable springs, respectively.
Independently from other hydrogeological and hydrological features of karst aquifers,
this observation may provide useful tools for water management purpose, especially
under droughts.
Ford & Williams (2007) provided a discussion on the applicability of the Darcys
law to karst and highlighted how a large part of the flow passes through the conduit
124 Karst without Boundaries
system, Thus, according to Mangin (1975), they assumed that the range of condi-
tion under which Darcys Law is valid is very restricted. The restriction has to refer
to different parts of the aquifer (spatial meaning), but also to the different recharge
conditions of the aquifer, which vary during the hydrological year (temporal meaning)
(Fiorillo, 2014). A Darcian flow prevalently occurs during the latest part of the reces-
sion, when the mean water flow velocity in the aquifer is the lowest with respect to all
previous stages. The Darcian hydraulic behaviour has to be connected to the drainage
of the water from the minor fissures, even if the conduit network provides the final
water transfer to spring (Fiorillo, 2014).
During the initial part of the recession, due to higher water flow velocites in the
conduit network, a different hydraulic flow type could occur. The initial and steepest
part of the recession can be modelled by a non-Darcian flow (Bailly-Comte et al.,
2010; Fiorillo, 2011; Malk & Vojtkov, 2012). The exchange of water between con-
duits and matrix occurs during floods (Haliban et al., 1998; Martin and Dean, 2001;
Bailly-Comte et al., 2010), and it is caused by the higher conduit head in the conduit
network than the matrix (Atkinson, 1977; Drogue, 1980; Bailly-Comte et al., 2010;
Fiorillo, 2011). This hydraulic behaviour depends on two main hydraulic heads which
characterise karst aquifers, and is caused by the two main permeability characteristics
of the medium: (i) the hydraulic head of the saturated zone (water table) and (ii) the
hydraulic head of the conduit network. Only during the long-term flow recession do
the two main different hydraulic heads tend to be the same, and they diverge after
recharge events, due to faster increase of the conduit hydraulic head.
Schematic illustrations highlight the main points of the aquifer hydrodynamics,
and help to associate the different parts of a hydrograph to specific hydraulic con-
ditions (Figure9.4). Non-allogenic recharge has been considered here, and a single
outlet (spring) drains the aquifer. In each sketch, the hydraulic head into the conduit
network is marked by the water level in the shafts; this level can be different from the
saturated aquifer zone, which is a function of the hydrologic condition of the aquifer.
Phase A shows the hydrological condition after a non-recharge period, character-
ised by a decreasing trend in the discharge, and the contribution from the vadose zone
can be neglected. In this phase the saturated zone is drained by the conduit network,
which has an hydraulic head similar or little lower than the saturated zone.
Phase B occurs after an intense rainfall event and is characterised by several rainy
hours or a few days of rainfall. The shafts and sinkholes allow surface runoff to
infiltrate and rapidly reach the saturated zone of the aquifer. The amount of rainfall
which causes the concentrated infiltration depends on the rainfall intensity and its
distribution over time, but is strongly controlled by the hydraulic characteristics of the
vadose zone, such as its thickness, the presence of shafts and the morphological char-
acteristics of the ground surface (presence of endorheic areas and swallow holes, slope
angle distribution, etc.). During phase B, shafts are temporarily filled with water, and
an increase of hydraulic head in the karst conduit network occurs, up to the spring
(Figure9.4). As the concentrated recharge varies according to the catchment fea-
tures, it provides a different water level for each shaft, which can cause a temporary
and locally different water flow direction into the saturated aquifer zone. The rapid
response of a spring to an intense rainfall event, with a typical peak in the spring
hydrograph, can be associated with the rise and lowering of the water level inside
conduits above the saturated zone terminating into the phreatic zone (Drogue, 1980;
Karst aquifer Aquiclude
a
LEGEND b WT, water table
A B C D E c
d HL, horizontal line
Q01 A)
Q =Q 0
Discharge, Q
1
t
Q011 WT
Q= HL Spring
Q 11
0 e
2 t
Q0111Q=Q 111e 3 t
lt
0
Fau
Time, t
B) rainfall
C)
Concentrated
Run off infiltration
diffuse
infiltration
Conduit network Conduit network
hydraulic head hydraulic head percolation
WT WT
HL Spring HL Spring
lt
lt
Fau
Fau
D) E)
Conduit network
hydraulic head
WT
WT
HL Spring HL Spring
lt
lt
Fau
Fau
Figure9.4 Spring hydrograph triggered by a storm pulse, where the different parts are associated with
sketches AE (modified from Fiorillo, 2014), showing the hydrodynamics in a karst aquifer.
Recharge occurs by diffuse percolation through the vadose zone, and in a concentrated
manner at points such as sinkholes/swallow holes and shafts; a basal spring drains the water
table by a conduit network. The hydrodynamics is shown by different sketches where the
slope of the water table is exaggerated. A, after a period of no recharge (a conduits and
shafts; b vadose zone; c and d saturated zone with different effective porosity). B, during a
period of strong recharge, concentrated infiltration causes the temporary filling of shafts,
later the aquifer water table rises as a consequence also of the diffuse infiltration. C, after a
period of strong recharge, water level in the shafts falls, and the water table could still rise.
D, lowering of the water table during a non-recharge phase; E, lowering of the water table
after a long period without recharge.
126 Karst without Boundaries
Bonacci & ivaljevic, 1993; Bonacci, 1995; Halihan et al., 1998). However, the very
high hydraulic conductivity of the conduit network can limit the rise (and lowering)
of the water level in conduits.
After the beginning of the increase in flow, a decrease of the chemical hardness
of the spring water is observed, which may mark a large part of the peak hydrograph
zone. Ashton (1966) provided a simple procedure to evaluate the volume of storage
in fully submerged conduits feeding a spring using discharge and the hardness of the
spring water. These observations indicate that the first part of the recession (the steep-
est part) discharges fresh water, which comes from the shafts and conduits, connected
to the springs by the conduit network.
The exchange of water between conduits and matrix occurs during floods (Martin
& Dean, 2001; Bailly-Comte et al., 2010), and it is caused by the higher conduit head
in the conduit network than the matrix (Atkinson, 1977; Drogue, 1980; Bailly-Comte
et al., 2010; Fiorillo, 2011). During phase B shafts and the conduit network feed the
water table as diffuse infiltration takes place as well. Under such hydraulic conditions,
spring discharge is connected to concentrated recharge processes and depends on the
hydraulic head in the conduit network. Some karst springs (for example, weakly
karstified systems) do not show this hydraulic behaviour, and their hydrograph shape
may appear completely smooth; in this case, spring discharge increases or decreases as
a function of long wet or dry periods, respectively (Fiorillo, 2011).
In heavily karstified systems, when recharge reduces or ends, a fall of discharge
can be observed at each spring. This process reflects rapid drainage of the conduit
network which and its temporarily filled shafts (Phase C). Phase C has a linear equa-
tion (which represents the nonlinear part in the semilogarithmic plot), but it may have
also another analytical form (Malk & Vojtkov, 2012). Mangin (1975) explained
this initial part of the recession as the influenced stage, characterised by the arrival (at
the water table) of the diffuse recharge from the unsaturated zone, which decreases
with time.
During phase C the hydraulic head in the conduit network decreases but it is still
higher than the water table. It may increase due to percolation from diffuse recharge.
When the water level in the shafts decrease and reach the water table level, the
hydraulic head in the conduit network reaches that of the matrix/minor fractures,
causing the drainage of the saturated zone by the conduit network, and phase D
begins. Many karst springs show an abrupt change of the hydrograph slope from
phase C to phase D; the gently sloped part of phase D highlights the higher water
volume stored in the saturated zone, which causes a lower decrease of the discharge
through time (Fiorillo, 2014). During this stage, the emptying processes cause the
lowering of the water table and involves water from minor conduits and the matrix,
where fluid viscosity and friction control the hydrograph recession shape. The con-
cave shape of this part of the hydrograph highlights, at least in part, the energy lost by
the water flow. Figure9.4 shows the recession during phase D by a single exponential
term, but the first part could still be influenced by the arrival of the diffuse infiltration,
and only later the exponential form could appear. Besides, following the theoretical
models (Rorabaugh, 1964; Brutsaert, 1994; Kovacs et al., 2005), this first part of
Phase D could be characterised by the sum of several exponentials.
Phase E refer to the hydrological condition after a long period without recharge,
during which the water table reaches the minimum height, following a different
Spring hydrograph recession: A brief review focused on karst aquifers 127
recession coefficient, a 3 (or a n). If the exponential form is expected, the changing of
the recession coefficient, from a 2 (phase D) to a3 (phase E) may depend on several
factors, and reflects the anisotropy of the aquifer or its non constant area during
drainage (Fiorillo, 2011).
9.4 CONCLUSION
REFERENCES
Ashton K. (1966) The analysis of flow data from karst drainage systems. Trans. Cave Res.
Group of Great Britain, 7, 161203.
Atkinson T.C. (1977) Diffuse flow and conduit flow in limestone terrain in Mendip Hills,
Somerset (Great Britain). Journal of Hydrology 35, 93100.
Baedke SJ, Krothe NC (2001). Derivation of effective hydraulic parameters of a karst aquifer
from discharge hydrograph analysis. Water Resources Research 37(1), 1319.
Bailly-Comte V., Martin J.B., Jourde H., Screaton E.J., Pistre S., Langston A. (2010) Water
exchange and pressure transfer between conduits and matrix and their influence on hydrody-
namics of two karst aquifers with sinking streams. Journal of Hydrology 386, 5566.
Bonacci O. (1995) Ground water behaviour in karst: example of the Ombla Spring (Croatia).
Journal of Hydrology 165, 113134.
Bonacci O., Zivaljevic R. (1993) Hydrological explanation of the flow in karst: example of the
Crnojevica spring. Journal of Hydrology 146, 405419.
Brutsaert W. (1994) The unit response of ground water outflow from a hill-slope Water
Resources Research 30(10), 27592763.
128 Karst without Boundaries
Dewandel B., Lachassagne P., Bakalowicz M., Weng Ph., Al-Malki A. (2003) Evaluation of
aquifer thickness by analysing recession hydrographs. Application to the Oman ophiolite
hard-rock aquifer. Journal of Hydrology 274, 248269.
Dreybrodt W., Romanov D., Kaufmann G. (2010) Evolution of caves in porous limestone by
mixing corrosion: a model approach. Geologia Croatica 63(2), 129135.
Drogue C. (1980) Essai didentification dun type de structure de magasins carbonats fis-
surs: application linterprtation de certains aspects du fonctionnement hydrogologique.
Mmoire hors srie de la Socit Gologique de France 11, 101108.
Eisenlohr, L., Kiraly, L., Bouzelboudjen, M., Rossier, I. (1997) A numerical simulation as a
tool for checking the interpretation of karst springs hydrographs. Journal of Hydrology 193,
306315.
Estrela T., Sahuquillo A. (1997) Modelling the response of a karstic spring at Arteta aquifer in
Spain. Ground Water 35(1), 1824.
Fiorillo F. (2011) Tank-reservoir emptying as a simulation of recession limb of karst spring
hydrographs. Hydrogeology Journal 19, 10091019.
Fiorillo F., Revellino P., Ventafridda G. (2012) Karst aquifer drainage during dry periods.
Journal of Cave and Karst Studies. 74(2), 148156.
Fiorillo F. (2014) The recession of spring hydrographs, focused on karst aquifers. Water
Resources Management, 28, 17811805.
Ford D. C., Ewers R. O. (1978). The development of limestone cave systems in the dimensions
of length and depth. Canadian Journal of Earth Science 15(11), 17831798.
Ford D., Williams P. (2007). Karst Hydrogeology and Geomorphology. Wiley, England, 562 pp.
Forkasiewicz, J., Paloc, H. (1967). Le regime de tarissement de la Foux-de-la-Vis [Analysis of
the recession period of the Foux-de-la-Vis spring]. Etude preliminaire. Chronique dHydro-
geologie, BRGM 3(10), 6173.
Gabrovek F., Dreybrodt W. (2010). Karstification in unconfined limestone aquifers by mixing
of phreatic water with surface water from a local input: a model. Journal of Hydrology
386(14), 130141.
Halihan T., Wicks C.M., Engeln J.F. (1998). Physical response of a karst drainage basin to flood
pulses: example of the Devils Icebox Cave system (Missouri, USA) Journal of Hydrology
204, 2436.
Kovcs A., Perrochet P. (2008). A quantitative approach to spring hydrograph decomposition.
Journal of Hydrology 352, 1629.
Kovcs A., Perrochet P., Kirly L., Jeannin P. (2005). A quantitative method for characterisa-
tion of karst aquifers based on the spring hydrograph analysis. Journal of Hydrology 303,
152164.
Jeannin P-Y., Sauter M. (1998). Analysis of karst hydrodynamic behaviour using global
approaches: e review. Bulletin dHydrogeologie 16, 3148.
Maillet E. (1905). Essais dHydraulique souterraine et fluviale [Underground and river hydrol-
ogy]. Hermann, Paris, 218 pp.
Malk P., Vojtkov S. (2012). Use of recession-curve analysis for estimation of karstification
degree and its application in assessing overflow/underflow conditions in closely spaced
karstic springs. Enveromental Earth Science 65(8), 22452257.
Martin, J.B., Dean, R.W. (2001) Exchange of water between conduits and matrix in the Floridan
aquifer. Chemical Geology 179(14), 145165.
Mangin A. (1975). Contribution ltude hydrodynamique des aquifres karstiques [A con-
tribution to the study of karst aquifer hydrodynamics]. 3me partie, Annales de Speleologie
30(1), 21124.
Milanovic P. (1976). Water regime in deep karst: case study of Ombla spring drainage area.
In: V. Yevjevich (ed.) Karst Hydrology and Water resources, vol. 1 Karst Hydrology, Water
resources Publications, Colorado, 165191.
Spring hydrograph recession: A brief review focused on karst aquifers 129
Nutbrown D.A., Dowing R.A. (1976) Normal-mode analysis of the structure of base flow
recession curves. Journal of Hydrology 30, 327340.
Rorabaugh M.I. (1964) Estimating changes in bank storage and groundwater contribution to
streamflow. International Association of Scientific Hydrology 63, 432441.
Sahuquillo A, Gmez-Hernndez, J.J. (2003) Comment on Derivation of effective hydraulic
parameters of a karst aquifer from discharge hydrograph analysis by SJ Baedke and NC
Krothe. Water Resources Research 39(6): 1152.
Stevanovic Z., Milanovic S., Ristic V. (2010) Supportive methods for assessing effective poros-
ity and regulating karst aquifers. Acta Carsologica 39(2), 301311.
Szilagyi J. (1999) On the use of semi-logarithmic plots for baseflow separation. Ground Water
37(5), 660662.
White W.B. (1988). Geomorphology and Hydrology of Karst Terrain. Oxford Press University,
pp. 464.
This page intentionally left blank
Chapter 10
Characterisation of selected
karst springs in Slovenia by
means of a time series analysis
Gregor Kovac ic 1 & Nataa Ravbar 2,3
1
University of Primorska, Faculty of Humanities and Science and
Research Centre, Koper, Slovenia
2
Karst Research Institute ZRC SAZU, Postojna, Slovenia
3
Urban Planning Institute of the Republic of Slovenia, Ljubljana,
Slovenia
ABSTRACT
The relation between discharge or water level, temperature and rainfall is evaluated for nine
karst springs/rivers using a daily time series analysis (autocorrelation and cross-correlation)
covering a 30-year period (19842013). The study was conducted in order to qualitatively
assess the applicability of long-term hydrological data set usage in time series analysis in order
to compare different karst springs and their catchments characteristics. The results reveal that
the storage capacity of larger systems and systems characterised by a more complex structure
is typically greater, although such systems should not necessarily be defined as poorly karsti-
fied. Factors influencing the obtained results are determined in the study. Application of longer
hydrological data sets is shown to provide valuable information on the hydrological properties
of springs, hydrodynamic behaviour and the hydraulic properties of the corresponding aqui-
fers. However, caution should be used when classifying karst systems into groups solely on the
basis of the results of a time series analysis.
10.1 INTRODUCTION
Univariate and bivariate time series analysis of hydrological data sets in the time
(autocorrelation and cross-correlation) and frequency (single spectral density and
cross-spectral density with corresponding amplitude, coherence, gain and phase func-
tions) domains have been widely used in karst hydrology. Jeannin & Sauter (1998)
argue that it is inappropriate to use these methods without good knowledge of the
investigated area to characterise the hydrogeological structure of karst aquifers, and
that the results obtained should be verified by deterministic models and/or direct obser-
vations. Despite this, the time series analysis methodology, first applied by Mangin
(1984) to study the input-output relationship in karst aquifers, has been further devel-
oped, improved and applied by many authors (e.g. Benavente etal., 1985; Padilla &
Pulido-Bosch, 1995; Angelini, 1997; Eisenlohr etal., 1997; Larocque etal., 1998;
Labat etal., 2000; Samani, 2001; Amraoui etal., 2003; Jukic & Denic-Jukic, 2004;
Mathevet etal., 2004; Rahnemaei etal., 2005; Valdes etal., 2005, 2006; Massei
etal., 2006; Panagopoulos & Lambrakis, 2006; Novel etal., 2007; Bailly-Comte
etal., 2008; Covington etal., 2009; Herman etal., 2009; Jemcov & Petric, 2009;
Fiorillo&Doglioni, 2010; Jemcov& Petric, 2010; Kovacic, 2010b; Terzic etal., 2012;
132 Karst without Boundaries
Delbart etal., 2014; Mayaud etal., 2014; Katsanou etal., 2015). All these studies of
various hydrological data sets (discharge, water level, precipitation, electrical conduc-
tivity, temperature, turbidity) have proved the usefulness of time series analysis as a
stand-alone or complementary method for the study of the functioning, hydrodynamic
behaviour and dynamics of karst springs and the corresponding aquifers.
Time series analysis has mostly been applied to periods covering i) one hydro-
logical year with daily or hourly time series (e.g. Larocque etal., 1998; Massei etal.,
2006; Herman etal., 2009; Kovacic, 2010b; Terzic etal., 2012) and ii) several suc-
cessive hydrological years with daily data (e.g. Benavente etal., 1985; Padilla &
Pulido-Bosch,1995; Angelini, 1997; Eisenlohr etal., 1997; Larocque etal., 1998;
Labat etal., 2000; Samani, 2001; Amraoui etal., 2003; Mathevet etal., 2004;
Rahnemaeietal., 2005; Panagopoulos & Lambrakis, 2006; Novel etal., 2007; Bailly-
Comte etal., 2008; Herman etal., 2009; Jemcov & Petric, 2009; Fiorillo & Doglioni,
2010; Jemcov& Petric, 2010; Kovacic, 2010b; Delbart etal., 2014; Katsanou etal.,
2015). Only a few studies (e.g. Valdes etal., 2005, 2006; Bailly-Comte etal., 2008;
Covington etal., 2009; Mayaud etal., 2014) have applied the method to a very short
or single-event timescale with hourly or even shorter time series. These studies have
demonstrated that single-event analysis provides relevant information on how a karst
system reacts to a single recharge event. However, long-term analysis provides infor-
mation about more average karst aquifer functioning (Panagopoulos & Lambrakis,
2006; Kovacic 2010b).
Herman etal. (2009), Delbart etal. (2014) and Kovacic (2010b) have shown
that the selection and duration of the reference period (one or more hydrological
years, parts of an individual hydrological year) within a time series analysis can
have strong effects on the results. For example, a comparative autocorrelation and
cross-correlation analysis performed for the Unica karst spring (Slovenia) in the five
successive hydrological years 19972002 revealed important differences in the cal-
culated memory effects (rk = 0.2) and the response times and values of maximum
precipitation-discharge cross-correlation coefficients (Kovacic, 2009). The charac-
teristic lags for the Unica spring for an individual hydrological year range from 17
to 86days, whereas the maximum cross-correlation coefficients vary between 0.21
and 0.33 with lags of 0 to 3 days (Kovacic, 2009). According to Mangins classifi-
cation (1984), the Unica spring could, therefore, be classified as a system with a low
(well-karstified) or very high memory effect (poorly karstified), which is rather con-
fusing. The study showed that the selection of the hydrological year (the structure of
the time series itself) could have significant effects on the results of a time series anal-
ysis (Kovacic, 2010b). Additionally, the study confirmed that the calculated memory
effects and response times to precipitation events also depend on the selection of the
beginning and the end of the hydrological year(s) in an investigated period.
Since in the literature time series analyses are only rarely employed for the char-
acterisation of karst aquifers for periods longer than 10 hydrological years (e.g.
Labat etal., 2000; Mathevet etal., 2004; Fiorillo & Doglioni, 2010), this chapter
represents a step forward, in that it is a comprehensive study of nine well-investigated
karst springs, the sinking rivers recharging these springs and the corresponding karst
aquifers in Slovenia. Time series analysis was performed on daily precipitation, dis-
charge, water level and temperature data sets for the 30-year period 19842013.
A30-year period was selected because this is a standard long-term reference period
Characterisation of selected karst springs in Slovenia 133
In karst hydrology, time series analysis of hydrological data requires the mathemati-
cal analysis of the response of a karst system to recharge (precipitation, concentrated
infiltration via ponors) and indirectly provides information regarding the structure
and functioning of karst aquifers (Box & Jenkins, 1970; Mangin, 1984; Box etal.,
1994). Univariate analysis reveals the structure of an individual time series, in terms
of either the time domain (autocorrelation) or the frequency domain (spectral den-
sity). The autocorrelation function, presented in a correlogram, quantifies the mem-
ory effect of the system, which is computed on the basis of the decorrelation lag time,
defined as the time at which the autocorrelation function attains a predetermined
value, usually 0.2 (Mangin, 1984; Benavente etal., 1985). The value of rk = 0.2 was
also used in this study. The memory effect reflects the inertia of the system. The slope
of the correlogram is determined by the response of the system to an event and gives
indirect information on the storage capacity of the system and its degree of karsti-
fication. Generally, a high memory effect indicates that the system is influenced by
an event for a long time, which often relates to a large storage capacity (Mangin,
1984). The correlogram of such karst aquifers will decrease slowly. Such systems
are characterised by a network of smaller fissures through which the water flows at
much lower velocities. In contrast, according to the literature (e.g. Larocque etal.,
1998; Panagopoulos & Lambrakis, 2006), a well-developed karst aquifer with larger
conduits and without a significant water storage capacity should correspond to a
low memory system, which, however, is a rather problematic statement. In general, a
correlogram of a well-developed karst system decreases steeply and quickly. Kovacic
(2010b) proved that large and well-developed karst aquifers such as the Malencica
and Unica karst springs systems can also have a large storage capacity, even though
the water flows through the system of unobstructed conduits at relatively high veloci-
ties (several tens of m/h; Gabrovek etal., 2010), and cannot be considered as poorly
karstified. In this regard caution is needed when interpreting the results on the basis of
the expected values. A spectral density function quantifies the regulation time, which
defines the duration of the influence of the input signal and gives an indication of the
length of the impulse response of the system (Larocque etal., 1998).
Cross-correlation and cross-spectral density functions imply the transformation
of input signals to output signals. In karst hydrology these are mostly discharge-
discharge, precipitation-discharge or water level-discharge relationships. The shape
of the cross-correlogram indicates the degree of karstification of a karst system.
The delay, which is the time lag between lag 0 and the lag of the maximum value
134 Karst without Boundaries
The catchment areas (relating to karst aquifers) of the karst springs (rivers) included
in this study are mostly developed in deep sequences of Mesozoic limestones and
dolomites, with the exception of the Pivka and Riana springs, where Palaeocene
Characterisation of selected karst springs in Slovenia 135
limestones also occur (Figure 10.1). The Riana, Malencica, Hubelj, Vipava and
Bistrica karst springs are tapped for drinking water supply.
Legend
Karst spring
Ponor
Gauging station
Precipitation station
Surface water
National boundary
Carbonate rocks
Intermittent lake
General groundwater
flow drainage
N
0 5 10 km Cartography:
N. Ravbar.
Figure 10.1 Location of investigated karst springs (rivers), precipitation and gauging stations.
136 Karst without Boundaries
Rovte
a
Vrhnika
ic
anica N
vsc
Ljubljanica Ljublj
tko
springs
Pe
Ljublja
jka
ten
Logatec polje Bistra
Ho
Hru
cica
Logas
sic
Men
a
isija
Pokojisce
Planina polje
Unica
Hasberg a
cic
Unica Malni r k nis
Malenscica Ce
Rakov Bl
a os
s cic Postojnska Skocjan cic
a
n o jama Kotlic i Rak
a Blo
towa N
rds V Postojna ke
ipava St
sprin Strzen r erovniscica
g z
en Z
Prestranek Dolenje
Jezero Cernica polje Lipsenjscica
Pivk
Pocek
Piv
Jav
ka
Ob Veliki
bas
Figure 10.2 Ljubljanica karst river basin (modified after ibrik etal., 1976; Kovacic, 2010a). Precipitation
and gauging stations included in the study are underlined.
karst waters from the Javorniki plateau, Loko Polje and the Bloke plateau. The major
sinking stream flowing into the polje is the Stren, which collects karst waters emerging
from numerous karst springs of varying abundance on the poljes eastern and southern
sides and partly on the western side at the foot of the Javorniki and Bloke plateaus
(Habic, 1976; Kovacic, 2010a). The right tributary of the Stren is the Cerknicica,
a surface river with a catchment of about 45 km2 (EARS, 2014a). The sum total of
inflows to the polje is 210240 m3/s; maximum runoff can surpass 90 m3/s (Kranjc,
1986; Kovacic, 2010a). As well as disappearing underground in the principal ponor
caves on the poljes north western margin (underground flow towards Rakov kocjan),
water from the polje sinks in numerous swallow holes. The water sinking through
Table 10.1 Key characteristics of the karst springs and watercourses in the study area (daily data for the period 19842013) with characteristic lag
times (days), where the autocorrelation coefficient exceeds the 0.2 value, and calculated regulation times. The locations of the measuring
sites are shown in Figures 10.1 and 10.2.
Characteristic
lag (days)
Spring or river Mean Min Max q5 q95 Qmin: Qmean: Qmax SD CV (%) (rk = 0.2) Treg
Malencica
Q (m3/s) 6.38 1.10 11.24 2.20 9.90 1: 5.80: 10.21 2.39 37.44 56 32
Unica
Q (m3/s) 20.25 1.08 89.22 2.29 62.10 1: 18.72: 82.46 19.96 98.57 47 39
T (C) 8.96 0.8 18.4 3.60 14.40 3.23 36.14 72 107
Stren
h (cm) 225.88 0 654 33 437 120.96 47.27 51 32
Pivka (Prestranek)
Q (m3/s) 2.49 0 43.35 0 11.10 3.95 158.61 40 13
Pivka (Postojnska jama)
h (cm) 123.55 7 829 34 220 60.24 48.76 39 17
Ljubljanica
Q (m3/s) 22.68 0.95 120.48 2.14 75.40 1: 23.77: 126.29 23.80 104.95 37 39
T (C) 9.88 3.7 14.8 6.1 13.2 2.30 23.25 78 58
Bistrica
Q (m3/s) 1.33 0.04 24.90 0.13 4.16 1: 30.11: 565.93 1.38 104.20 45 25
Vipava
Q (m3/s) 6.25 0.73 81.42 1.11 21.90 1: 8.59: 112.00 7.75 124.09 7 7
T (C) 10.35 8.0 15.0 8.8 13.0 1.28 12.38 r(999) = 0.385
Hubelj
Q (m3/s) 2.76 0.18 51.26 0.34 9.38 1: 15.77: 229.94 3.67 132.79 12 25
T (C) 8.53 5.0 13.3 7.3 10.3 0.92 10.78 72
Riana
Q (m3/s) 3.37 0.03 63.16 0.13 15.50 1: 112.3: 2105.27 5.48 162.60 15 22
138 Karst without Boundaries
these swallow holes re-emerges in the Ljubljanica springs (Behrens etal., 1976). Within
the Rakov kocjan karst depression, waters from Cerknica Polje and, in part, from the
Javorniki plateau flow into the Rak, a surface stream. When the Rak finally disappears
underground, it flows towards the Unica and Malencica springs.
To the north west is Planina Polje (the outflow of the Malencica and Unica
springs), an important confluence of karst waters from three contributing sub-
catchments of the Ljubljanica basin: Cerknica Polje, the Pivka basin, and the Snenik
and Javorniki karst plateaus. The Malencica karst spring is a tributary of the more
abundant Unica river. Their catchment areas almost completely overlap, with the
exception of the Lower Pivka basin area (the Pivka with the Nanocica), which is a
part of the Unicas catchment but not part of the Malencicas catchment. In compar-
ison to the Unica spring, the portion of autogenic recharge in the Malencica spring
is greater. The Unica river sinks on the eastern and northern rim of Planina Polje
and flows directly towards the Ljubljanica karst springs. Surface drainage occurs in
the Lower Pivka basin (Eocene flysch) and in the Cerknicica and Blocica drainage
basins (Triassic clastic formations). Alluvial deposits from the Quaternary period
are found along the Cerknicica, Blocica and Pivka rivers and at the bottom of the
higher-lying poljes.
According to the results of the water balance analysis for the 30-year observa-
tion period (19712000), the Malencica spring has a catchment area of 200km2
(Kovacic, 2010b). However, tracer test results conducted in the past show that its
catchment is much larger and extends over 726km2 (EARS, 2009). The catchment
area of the Unica is estimated at 800 km2 (ibrik etal., 1976). The mean discharge
of the Malencica and Unica springs in the reference period 19842013 is 6.38 and
20.25 m3/s respectively (EARS, 2014a). The ratio between minimum, mean and max-
imum discharge of the Malencica spring is 1: 5.8: 10.2, with a minimum discharge
of 1.10 and a maximum of 11.24 m3/s. Minimum discharge of the Unica spring in
the reference period is 1.08 while the maximum is 89.22 m3/s (EARS, 2014a). The
ratio between minimum, mean and maximum discharge is 1: 18.7: 82.5. When water
levels are low, the Malencica contributes almost all of the water in the Unica. Mean
annual precipitation in the common catchment area over the long-term period is
estimated at 1780 mm; mean annual evapotranspiration is 720 mm and mean annual
run-off is 1060 mm, with a runoff coefficient of around 60% (Kovacic, 2010b).
The Pivka river basin has a dual hydrogeological structure. The upper part of
the basin (Prestranek gauging station), which also drains the western parts of the
surrounding Javorniki and Snenik karst plateaus, is in limestones, while the lower
part is in almost impermeable Eocene flysch. The Pivka sinks in the world-famous
Postojna Cave and flows towards Planina Polje as the Pivka branch of Planina Cave.
The cave is a spring of the Unica river, where two subsurface river channels (the
Rak and Pivka branches) meet. The Rak branch drains the waters from the Rakov
kocjan karst depression and the Javorniki plateau. The catchment area of the Pivka
river at its ponor is estimated at 300 km2 (ibrik etal., 1976). In the reference period
19842013 the discharges of the Pivka (Prestranek gauging station) ranged between 0
and 43.35 m3/s, with a mean of 2.49 m3/s (EARS, 2014a). When water levels are low,
the Pivka dries up completely in the upper part of the basin.
North of Planina Polje is Logatec Polje with several small sinking streams
(Figure10.2). Together with the Logatec and Rovte plateaus, this part of the
Ljubljanica catchment covers about 100 km2 (Habic, 1976). The area is mainly built
Characterisation of selected karst springs in Slovenia 139
The mean discharge of the Riana spring in the period 19842013 is 3.37m3/s,
with a minimum of 0.03 and a maximum of 36.16 m3/s (EARS, 2014a). The ratio
between minimum, mean and maximum discharge is 1: 112.3: 2105.3. If in the calcu-
lation the amount of water abstracted for water supply is included (200 l/s), the ratio
changes to 1: 15.5: 275.5. Mean annual precipitation (19842013) in the catchment
(Podgrad precipitation station) is 1465 mm (EARS, 2014b). The recharge area of
the Riana spring covers an area of 245 km2. The spring discharges at an elevation
of 70m asl, while the recharge area is at 500 to 1100 m asl. It is predominantly
composed of carbonate rocks (Cretaceous and Palaeocene limestones). Several small
surface streams from the Eocene flysch of the Brkini hills sink underground and addi-
tionally recharge the karst aquifer (Krivic etal., 1987; Krivic etal., 1989).
1.0
Unica
Malenscica
Pivka Prestranek
0.8 Pivka Postojnska. h
Strzen h
Ljubljanica
0.6
0.4
r (k)
0.2
0.0
0.2
0 20 40 60 80 100 120
k (lag in days)
Figure 10.3 Autocorrelation functions of daily discharges for the Unica, Malencica, Ljubljanica, Stren
(water level) and Pivka (water levels in its middle course and at its ponor).
water levels of the Stren and the Malencica in one hydrological year were identical;
both autocorrelograms reached the rk = 0.2 value after 32 days.
Only the Pivka spring has two gauging stations that show a good correlation
(Spearmans rho = 0.86). The upper and lower courses of the Pivka show slightly dif-
ferent results in the time series analysis. The upper part of the basin is characterised
by a shallow karst aquifer with relatively fast recharge at high water levels (floods),
but is also subject to fast subsurface drainage towards the Planina Polje springs when
the water level is low. Unlike the lower course, the upper course of the Pivka at the
Prestranek gauging station usually dries up completely. The autocorrelation of the
Pivka (Prestranek) discharges exceeds the characteristic lag after 40 days and has a
regulation time of 13 days, showing that the storage capacity of the upper part of
the Pivka basin is small to medium. Data for the levels of the Pivka at the Postojna
Cave ponor show similar values: the characteristic lag is reached after 39 days and
the regulation time is 17 days which, surprisingly, is four days longer than at the
Prestranek gauging station. In its lower course, the Pivka is highly influenced by the
surface-flowing Nanocica, which ought to considerably decrease the memory effect
of the Pivka at its ponor (Postojnska jama) and shorten the regulation time. This,
however, is not the case, owing to: i) floods that occur outside the Pivka ponor when
the amount of incoming water exceeds the swallow capacity of the ponor, which
results in a prolongation of the water level time series persistence and ii) the use of
a water level hydrological time series in the analysis, which generally show smaller
variability (CV = 49%; Table10.1) than discharges. The influence of the Nanocica is
apparent in a CV for the temperature of the Pivka that is higher than that for any of
the other springs in the study (63%). The temperature characteristic lag at the Pivka
ponor is 77 days. The highest value of rxy between the temperatures of the Pivka ponor
and the Unica is reached at k= 14 (rxy = 0.74), indicating that the temperatures of
the Unica react faster.
142 Karst without Boundaries
The most abundant of the karst springs in the Ljubljanica river basin is the Unica.
Its memory effect is 47 days (Figure 10.3), which corresponds to a medium to high
storage capacity. Surprisingly, and for an unexplained reason, the regulation time of
the spring (39 days) exceeds that of the Malencica by seven days and indicates a
longer impulse response, which is exactly the opposite of the results calculated in the
study conducted by Kovacic (2009). The daily temperature autocorrelation function
of the spring shows a memory effect of 72 days, meaning that the impulse response
of the system to temperature is long. This has also been observed in the study area
by means of short-term time series analysis (Kogovek & Petric, 2010). Temperature
commonly shows a longer memory effect than other hydrological time series (dis-
charge, water level, electrical conductivity, turbidity), which corresponds to the low
variability of temperatures (Table 10.1). At the same time, a temperature time series
provides less information about the structure of the karst system, since it is not a con-
servative tracer (Larocque etal., 1998; Kovacic, 2009).
As we have already mentioned, the characteristics of the discharges of the
Malencica (Malni gauging station) are reflected in those of the Unica (Hasberg gaug-
ing station); the Spearmans rho between the stations is 0.96. A minor part of the
recharge of the Unica spring comes from the Pivka. The value of the same correlation
coefficient between water levels at the Pivka ponor and discharges of the Unica is 0.93.
The Spearmans rho correlation between the Unica and the Ljubljanica (0.94) also
confirms the significant contribution of the Unica to the recharge of the Ljubljanica
springs; their mean discharge exceeds the Unicas by just 12%. The Ljubljanica springs
have the same regulation time of 39 days as the Unica, showing that the length of the
impulse response of the system is medium. The memory effect of the springs is, how-
ever, slightly lower (37 days), which still indicates the medium storage capacity of the
system, although the latter value is the lowest among all the gauging stations studied
in the Ljubljanica river basin. One of the reasons for this could be the relatively high
outflow capacity of the Ljubljanica permanent and intermittent karst springs, which
can exceed 110 m3/s, meaning that the system could be subject to rapid drainage. The
Ljubljanica karst springs (CV = 105%) are also characterised by greater discharge
variability in comparison to the Unica (CV = 99%). Another reason could be the
influence of the Bela surface stream during intense precipitation. The latter collects
water from partly karstified dolomite and flows into the Ljubljanica upstream of the
gauging station (Habic, 1976). The temperature memory effect of the Ljubljanica
spring (78 days) is similar to that of the Unica. The cross-correlation temperature
function between the Unica and Ljubljanica shows a clear asymmetry towards posi-
tive k values, reaching the maximum value at k = +10 (0.872) and once more indicat-
ing the significant influence of the Unica on the Ljubljanica springs.
The autocorrelation function of the Bistrica spring, recharged only by autogenous
precipitation from the Snenik plateau, shows that the memory effect of the spring is
medium (45 days), while the regulation time of 25 days indicates the relatively short
duration of the influence of the input signals (Figure 10.4). It is uncertain whether
the memory effect also reflects the medium storage capacity of the spring or can be
attributed at least to some extent to the small total outflow capacity of the springs.
Unlike in the Ljubljanica river basin, where some large spring caves are present, there
are no known caves in the outflow zone of the Bistrica springs (Cave Registry, 2014).
The Hubelj karst spring drains the deeply karstified Trnovski Gozd plateau. The low
memory effect of the spring (12 days) corresponds to the small storage capacity of the
Characterisation of selected karst springs in Slovenia 143
1.0
Bistrica
Vipava
0.8 Hubelj
Rizana
0.6
0.4
r (k)
0.2
0.0
0.2
0 20 40 60 80 100 120
k (lag in days)
Figure 10.4 Autocorrelation functions of daily discharges for the Bistrica, Vipava, Hubelj and Riana
karst springs.
system and indicates the very rapid flow of infiltrated precipitation water through the
system of a hydraulically connected network of fissures and conduits in its vadose and
phreatic zones towards the spring. Over a shorter period, Kranjc (1997) calculated the
memory effect of the spring to be 10 days. The high value of the springs discharge CV
(133%) and groundwater level fluctuations of several tens of metres show that the out-
flow capacity of the spring is not limited and the spring can drain large amounts of water
in a short period of time. Surprisingly, the regulation time (25 days) is not completely in
accordance with the calculated memory effect of the spring, but again it implies the small
storage capacity and short impulse response of the spring. In terms of the temperature
autocorrelation function, the spring shows similar values to the Unica and Ljubljanica
springs (72 days). It also has the smallest variability (CV= 11%) of the temperature time
series in this study. The regulation time of the daily temperatures could not be calculated.
The highest discharge variability among all the spring studied is that of the Riana
karst spring (CV = 163%). The memory effect of the spring is low (15 days) which
corresponds to the small storage capacity of the aquifer, while the regulation time
(22days) indicates that the length of the impulse response of the system is short.
Among all the springs studied, the Vipava spring has the lowest memory effect
(7 days) which indicates a well-drained karst aquifer with larger and well-connected
conduits that prevent significant water storage. For a much shorter period, Jemcov &
Petric (2009) calculated a memory effect of 6 days. Although the Vipava springs are
also recharged by sinking streams, the importance of the latter is insignificant when
compared to autogenic recharge by precipitation water. The hydraulic conductivity of
the system of active fissures and conduits within the vadose zone of the Vipava aquifer
must be quite high to immediately transport the pressure pulse of infiltrated precipi-
tation water from the Nanos plateau towards the springs. The regulation time of the
spring is 7 days, compared to 14 days (Jemcov & Petric, 2009) and 12 days (Kranjc,
1997). The temperature autocorrelation function decreases slowly and shows values
above rk = 0.2 at lags exceeding 999 days.
144 Karst without Boundaries
The memory effects (rk = 0.2) of the various springs show no significant (p<
0.05) correlation (Spearmans rho = 0.32) with regulation times, while they do show
a weak positive (though not statistically significant) correlation with mean discharges
(Spearmans rho = 0.50). Generally, this means that more abundant karst springs with
larger catchments show higher memory effects, although these aquifers cannot as a
rule be considered poorly karstified. Interestingly, the values of the springs coefficient
of variation show a relatively high and statistically significant negative correlation
with the memory effects of the springs (Spearmans rho = -0.82), meaning that the
greater the variability of the springs discharges, the shorter the memory effects of the
springs.
Jurce Postojna Pokojice Podgrad Podkraj Stren Pivka (river) Pivka (ponor) Unica
Spring or river (P) (P) (P) (P) (P) (h) (Q) (h) (Q)
Malencica
Max rxy(k) 0.194 (1) 0.907 (-1) 0.692 (0)
COxy 0.519 0.508 0.417
MD 5.92 0.72 0.93
CUFF 16.89 9.24 17.17
Unica
Max rxy(k) 0.266 (1) 0.293 (1) 0.806 (-2) 0.906 (0) 0.756 (0)
COxy 0.577 0.627 0.529 0.534 0.559
MD 1.30 0.48 6.60 1.50 5.50
CUFF 5.25 3.18 5.60 7.95 3.32
Stren (h)
Max rxy(k) 0.191 (7)
COxy 0.538
MD 1.53
CUFF 3.44
Pivka (river)
Max rxy(k) 0.249 (2)
COxy 0.482
MD 0.47
CUFF 11.46
Pivka (ponor) (h)
Max rxy(k) 0.384 (1) 0.436 (0) 0.756 (0)
COxy 0.587 0.631 0.569
MD 1.85 3.64 2.82
CUFF 2.32 3.16 5.55
(Continued)
Table 10.2 (Continued).
Jurce Postojna Pokojice Podgrad Podkraj Stren Pivka (river) Pivka (ponor) Unica
Spring or river (P) (P) (P) (P) (P) (h) (Q) (h) (Q)
Ljubljanica
Max rxy(k) 0.392 (1) 0.355 (1) 0.934 (0)
COxy 0.631 0.636 0.703
MD 2.74 1.32 0.66
CUFF 2.91 3.00 3.27
Bistrica
Max rxy(k) 0.278 (2)
COxy 0.552
MD 5.98
CUFF 17.05
Vipava
Max rxy(k) 0.679 (0)
COxy 0.722
MD 0.49
CUFF 3.02
Hubelj
Max rxy(k) 0.634 (0)
COxy 0.634
MD 0.61
CUFF 9.51
Riana
Max rxy(k) 0.538 (0)
COxy 0.645
MD 1.16
CUFF 4.76
Characterisation of selected karst springs in Slovenia 147
1.0
Jursce Malenscica
Jursce Pivka Prestranek
Jursce Unica
0.8 Jursce Strzen (h)
Postojna Pivka Postojnska j. (h)
Pokojisce Ljubljanica
0.6
rxy (k)
0.4
0.2
0.0
40 20 0 20 40 60 80
k (lag in days)
Figure 10.5 Cross-correlation functions of discharges and daily precipitation series as inputs for the
Unica, Malencica, Ljubljanica, Stren (water level) and Pivka (water levels in its middle
course and at its ponor).
range from 2.32 days (Pivka, Postojnska jama) and 17.05 days (Bistrica), which points
to evident differences in the abilities of the springs in question to filter precipitation
events of varying duration.
Discharge (water level)-discharge (water level) cross-correlation coefficients
show higher values then precipitation-discharge cross-correlation values, and range
from 0.69 to 0.93 with delays of between -2 and 0 days (Figures 10.8 and 10.9). The
negative time lags (Stren-Malencica and Stren-Unica) correspond to the faster
response to recharge of the gauging station positioned downstream in comparison
to the upstream station. The highest cross-correlation coefficient is calculated for
the Unica-Ljubljanica springs (rxy = 0.93). The cross-correlogram shows a slight
asymmetry to the negative values, but an instant response to the recharge of both
springs. The value of the average coherence function (0.70), the mean delay (0.66
days) and the cut-off frequency (3.27 days) all imply well-developed underground
connections and a similar response of the springs to the more or less homogeneous
recharge of karst aquifers by precipitation in the catchments. The second highest
cross-correlation coefficient is calculated for the Stren-Malencica pair of variables
(rxy = 0.91). The high correlation value of the almost symmetrical function reflects
the fact that both time series have a similar structure. The discharge variability is
small, whereas the water levels of the Stren show a slow increase once the water in
the river channel starts to overflow and induces the appearance of Lake Cerknica
(Kovacic, 2010b).
The Pivka (Prestranek gauging station) is not directly connected to the Malencica
and Unica springs, but the cross-correlation coefficient shows how the springs react to
recharge from the upper part of the Pivka basin. Due to the complexity of its aquifer,
the Malencica spring shows much lower cross-correlation coefficient and coherence
function values than the Unica. However, the cross-correlation coefficient between the
148 Karst without Boundaries
Vrhnika
a
ic
nica N
Ljublja
vsc
tko
Pe
jka
0.3
ten
6(
Ho Logatec polje
1)
0.6
3
) 0.70
cica
0.6
Logas
4
(1)
0.93 (0
9
0.3
Un Pokojisce
ica
Planina polje
1( )0 os
1
)0 .53 cic
Postojnska a
(1)
.51 Dolenje
0.44 (0) 0.63 jama Jezero
0.29
a
cic Postojna
0.19 (1
s Str
a no 2 ze
N (0) 0.4 n
0.69
0.27
Prestranek
0.3
Cernica polje
(1)
8(
Legend
0.58
4
1)
Losko polje
0.5
0.5
Ob
9
5
9(
sinking stream (2
)0 rh
0.1
Watershed .48
Polje
Important karst spring Jursce
Piv
Gauging station
ka
water levels of the Pivka at its ponor and the discharges of the Unica is not high either
(0.76), which means that the input signal of the Pivka ponor is significantly trans-
formed on its way to the Unica spring. The influence of the Nanocica on the Pivka
at its ponor is reflected in the low value of the cross-correlation coefficient (0.76)
between the Prestranek and Postojnska jama gauging stations.
In comparison to the calculated precipitation-spring mean delays are much shorter
(ranging from 0.66 to 6.6 days) between sinking streams and springs, while the values
of the cut-off frequencies are very similar (Table 10.2).
Calculations of Spearmans rho show that a statistically significant (p < 0.05) neg-
ative correlation exists between the calculated memory effects of the springs (rivers)
Characterisation of selected karst springs in Slovenia 149
1.0
Podkraj Vipava
Podkraj Hubelj
Podgrad Bistrica
0.8 Podgrad Rizana
0.6
rxy (k)
0.4
0.2
0.0
40 20 0 20 40 60 80
k (lag in days)
Figure 10.7 Cross-correlation functions of discharges and daily precipitation series as inputs for the
Bistrica,Vipava, Hubelj and Riana karst springs.
1.0
Strzen (h) Malenscica
Strzen (h) Unica
0.8 Unica Ljubljanica
0.6
0.4
rxy (k)
0.2
0.0
0.2
100 80 60 40 20 0 20 40 60 80 100
k (lag in days)
Figure 10.8 Cross-correlation functions of discharges (water levels) as inputs and discharges as out-
puts for the Unica, Malencica and Ljubljanica karst springs.
1.0
Pivka Prestranek Malenscica
Pivka Prestranek Pivka Postojnska j. (h)
Pivka Postojnska j. (h) Unica
0.8 Pivka Prestranek Unica
0.6
0.4
rxy (k)
0.2
0.0
0.2
100 80 60 40 20 0 20 40 60 80 100
k (lag in days)
Figure 10.9 Cross-correlation functions of discharges (water levels) as inputs and discharges (water
levels) as outputs for the Unica and Malencica karst springs and for the Pivka river at its ponor.
correlation with the mean discharges (0.66), meaning that more abundant karst
springs filter longer-lasting precipitation events.
Autocorrelation Q
High MEMORY EFFECT Low
Malenscica
Strz en
Unica
Cross-correlation PQ
CROSS-CORR. COEFF.
RESPONSE TIME
Riz ana
Hubelj
Vipava
Short
High
well
karstified
Figure 10.10 Classification of the studied springs into groups according to the results of the time
series analysis. Binary karst systems are marked in bold.
sets (more successive hydrological years), since they include more information and
reflect more average conditions. Finally, the results of the time series analysis should
be interpreted together with the results of other methods (e.g. speleological investiga-
tions, tracer tests, modelling, etc.) used in karst hydrology in order to avoid bias in
conclusions.
REFERENCES
Amraoui F., Razack M., Bouchaou L. (2003) Turbidity dynamics in karstic systems. Example
of Ribaa and Bittit springs in the Middle Atlas (Morocco). Hydrological Sciences Journal-
Journal Des Sciences Hydrologiques 48(6), 971984.
Angelini P. (1997) Correlation and spectral analysis of two hydrogeological systems in Central
Italy. Hydrological Science - Journal-des Sciences Hydrologiques 42(3), 425438.
Bailly-Comte, V., Jourde, H., Roesch, A., Pistre, S., Batiot-Guilhe, C. (2008) Time series anal-
yses for karst/river interactions assessment: case of the Coulazou River (southern France).
Journal of Hydrology 349(12), 98114.
Benavente J., Pulido-Bosch A., Mangin A. (1985) Application of correlation and spectral proce-
dures to the study of the discharge in a karstic system (Eastern Spain). In: Gunay, G., Johnson
A.I. (editors) Karst water resources Proceedings of the Ankara-Antalya Symposium, July
1985, 6775.
Behrens H., Zupan M., Zupan M. (1976) Tracing with fluorescent tracers. In: Gospodaric
R., Habic P. (editors) Underground water tracing: Investigations in Slovenia 19721975.
Third International Symposium of Underground Water Tracing (3. SUWT) Ljubljana-Bled,
139164.
Box G.E.P., Jenkins G.M. (1970) Time series analysis: Forcasting and control. Holden Day, San
Francisco.
Box G.E.P., Jenkins G.M., Reinsel C. (1994) Time series analysis: forecasting and control,
3rded. Prentice Hall, New Jersey.
Cave Registry (2014) Speleological Association of Slovenia and Karst Research Institute
ZRC SAZU.
Covington, M.D., Wicks, C.M., Saar, M.O. (2009) A dimensionless number describing the
effects of recharge and geometry on discharge from simple karstic aquifers. Water Resources
45(11), W11410.
Delbart C., Valdes D., Barbecot F., Tognelli F., Richon P., Cauchoux L. (2014) Temporal
variability of karst aquifer response time established by sliding-windows cross-correlation
method. Journal of Hydrology 511, 580588.
Eisenlohr L., Bouzelboudjen M., Kiraly L., Rossier Y. (1997) Numerical versus statistical model-
ling of natural response of a karst hydrogeological system. Journal of Hydrology 202, 244262.
Environmental agency of the Republic of Slovenia (EARS), Digital data on water (2009) URL:
http://gis.arso.gov.si; last visited 1/7/2009.
Environmental agency of the Republic of Slovenia (EARS) (2014a) Daily discharges and water
levels at the Malni (the Malencica), Hasberg (the Unica), Postojna Cave (the Pivka ponor),
Prestranek (the Pivka), Dolenje Jezero (the Stren), Vipava I (the Vipava), Ajdovcina I (the
Hubelj), Kubed II (the Riana), Ilirska Bistrica (the Bistrica) and Vrhnika (the Ljubljanica)
gauging stations for the period 19842013. The archive of the Environmental agency of the
Republic of Slovenia, Ljubljana.
Environmental agency of the Republic of Slovenia (EARS) (2014b) Daily precipitation amounts
at the Podkraj, Podgrad, Jurce, Postojna and Pokojice rain gauge stations for the period
19842013. The archive of the Environmental agency of the Republic of Slovenia, Ljubljana.
Characterisation of selected karst springs in Slovenia 153
Fiorillo F., Doglioni A. (2010) The relation between karst spring discharge and rainfall by
cross-correlation analysis (Campania, southern Italy). Hydrogeology Journal 18, 18811895.
Gabrovek F., Kogovek J., Kovacic G., Petric M., Ravbar N., Turk J. (2010) Recent results
of tracer tests in the catchment of the Unica River (SW Slovenia). Acta Carsologica 39(1),
2737.
Habic P. (1976) Geomorphologic and hydrographic characteristics. Investigations in
Ljubljanica river basin, Description of investigated area. In: Gospodaric R., Habic P. (editors)
Underground water tracing: Investigations in Slovenia 19721975. Third International
Symposium of Underground Water Tracing (3. SUWT) Ljubljana-Bled, 1227.
Habic P. (1989) Kraka bifurkacija na jadransko crnomorskem razvodju. Acta carsologica 18,
235264.
Herman E.K., Toran L., White W.B. (2009) Quantifying the place of karst aquifers in the
groundwater to surface water continuum: a time series analysis study of storm behaviour in
Pennsylvania water resources. Journal of Hydrology 376, 307317.
Jemcov I., Petric M. (2009) Measured precipitation vs. effective infiltration and their influence
on the assessment of karst systems based on results of the time series analysis. Journal of
Hydrology 379(34), 304314.
Jemcov I., Petric M. (2010) Time series analysis, modelling and assessment of optimal exploita-
tion of the Nemanja karst springs, Serbia. Acta Carsologica 39(2), 187200.
Jane J., car J., Habic P., Podobnik R. (1997) Vodno bogastvo Visokega krasa. Geologija d.o.o.,
Idrija.
Jeannin P. Y., Sauter M. (1998) Analysis of karst hydrodynamic behaviour using global
approach: a review. In: Jeannin, P. Y., Sauter M. (editors) Modelling in karst systems. Bulletin
dHydrogeologie 16, 3148.
Jukic D., Denic-Jukic V. (2004) A frequency domain approach to groundwater recharge esti-
mation in karst. Journal of Hydrology 289, 95110.
Katsanou K., Lambrakis N., Tayfur G., Baba A. (2015) Describing the Karst Evolution by the
Exploitation of Hydrologic Time-Series Data. Water Resources Management 29, 31313147.
Kogovek J. (1999) Nova spoznanja o podzemnem pretakanju vode v severnem delu Javornikov
(Visoki kras). Acta Carsologica 28(1), 161200.
Kogovek J., Knez M., Mihevc A., Petric M., Slabe T., ebela S. (1999) Military training area in
Kras (Slovenia). Environmental Geology 38(1), 6976.
Kogovek J., Petric M. (2010) Water temperature as a natural tracer a case study of the
Malencica karst spring (SW Slovenia). Geologia Croatica 63(2), 171177.
Kovacic G. (2003): Kraki izviri Bistrice (JZ Slovenija). Annales: Series historia naturalis 13(1),
111120.
Kovacic G. (2009) Hydrology of the Malencica karst spring and its catchment. PhD thesis,
University of Primorska, Koper.
Kovacic G. (2010a) An attempt towards an assessment of the Cerknica Polje water balance.
Acta Carsologica 39(1), 3950.
Kovacic G. (2010b) Hydrogeological study of the Malencica karst spring (SW Slovenia) by
means of a time series analysis. Acta Carsologica 39(2), 201215.
Kranjc A. (1986) Cerkniko jezero in njegove poplave. Geografski zbornik 25, 73123.
Kranjc A. (editor) (1997) Karst hydrogeological investigations in south-western Slovenia. Acta
Carsologica 26(1), 1388.
Krivic P., Bricelj M., Triic N., Zupan M. (1987) Sledenje podzemnih vod v zaledju Riane.
Acta Carsologica 16, 83104.
Krivic P., Bricelj M., Zupan M. (1989) Podzemne vodne zveze na podrocju Cicarije in osrednjega
dela Istre (Slovenija, Hrvatska, NW Jugoslavija). Acta Carsologica 18, 265295.
Labat D., Ababou R., Mangin A. (2000) Rainfall-runoff relations for karstic springs. Part I:
convolution and spectral analyses. Journal of Hydrology 238, 123148.
154 Karst without Boundaries
Larocque M., Mangin A., Razack M., Banton O. (1998) Contribution of correlation and
spectral analyses to the regional study of a large karst aquifer (Charente, France). Journal of
Hydrology 205, 217231.
Mangin A. (1984) Pour une meilleure connaissance des systmes hydrologiques partir des
analyses corrlatoire et spectrale. Journal of Hydrology 67, 2543.
Massei N., Dupont J.P., Mahler B.J., Laignel B., Fournier M., Valdes D., Ogier S. (2006)
Investigating transport properties and turbidity dynamics of a karst aquifer using correlation,
spectral, and wavelet analyses. Journal of Hydrology 329, 244257.
Mathevet T., Lepiller M., Mangin A. (2004) Application of time-series analyses to the hydro-
logical functioning of an Alpine karstic system: the case of Bange-LEau-Morte. Hydrology
and Earth System Sciences 8, 10511064.
Mayaud C., Wagner t., Benischke R., Birk S. (2014) Single event time series analysis in a binary
karst catchment evaluated using a groundwater model (Lurbach system, Austria). Journal of
Hydrology 511, 628639.
Novel, J.P., Dimadi A., Zervopoulou A., Bakalowicz M. (2007) The Aggitis karst system,
Eastern Macedonia, Greece: Hydrologic functioning and development of the karst structure.
Journal of Hydrology 334, 477492.
Padilla A., Pulido-Bosch A. (1995) Study of hydrographs of karstic aquifers by means of cor-
relation and cross-spectral analysis (France, Spain). Journal of Hydrology 168, 7389.
Panagopoulos G., Lambrakis N. (2006) The contribution of time series analysis to the study
of the hydrodynamic characteristics of the karst systems: Application on two typical karst
aquifers of Greece (Trifilia, Almyros Crete). Journal of Hydrology 329, 368376.
Rahnemaei M., Zare M., Nematollahi A.R., Sedhi H. (2005) Application of spectral analysis of
daily water level and spring discharge hydrographs data for comparing physical characteris-
tics of karst aquifers. Journal of Hydrology 311, 106116.
Samani N. (2001) Response of karst aquifers to rainfall and evaporation, Maharlu basin, Iran.
Journal of Cave and Karst Studies 63( 1), 3340.
Terzic J., Stroj A., Frangen T. (2012) Hydrogeological investigation of karst system properties
by common use of diverse methods: a case study of Licka Jesenica springs in Dinaric karst of
Croatia. Hydrological Processes 26(21), 33023311.
Turk J., Malard A., Jeannin P-Y., Vouillamoz J., Masini J., Petric M., Gabrovek F., Ravbar N.,
Slabe T. (2013) Interpretation of hydrogeological functioning of a high karst plateau using
the KARSYS approach: the case of Trnovsko-Banjka planota (Slovenia). Acta Carsologica
42(1), 6174.
Valdes, D., Dupont, J.P., Massei, N., Laignel, B., Rodet, J. (2005) Analysis of karst hydrody-
namics through comparison of dissolved and suspended solids transport. C. R. Geoscience
337, 13651374.
Valdes, D., Dupont, J.P., Massei, N., Laignel, B., Rodet, J. (2006) Investigation of karst hydro-
dynamics and organization using autocorrelation and T-curves. Journal of Hydrology 329,
432443.
ibrik K., Lewicki F., Picinin A. (1976): Hydrologic investigations. Investigations in Ljubljanica
river basin, Investigations before the tracing test 19721975. In: Gospodaric R., Habic P.
(editors) Underground water tracing: Investigations in Slovenia 19721975. Third
International Symposium of Underground Water Tracing (3. SUWT) Ljubljana-Bled, 156.
Chapter 11
ABSTRACT
Alpine and Mediterranean karst aquifers are frequently associated with Upper Triassic (Keuper)
deposits which generally constitute the substratum or impervious base of the system given their
predominantly clayey nature. When the groundwater stored in the carbonate aquifer comes
into direct contact with the Triassic formation, the result is a modification of the chemical com-
position of the water due to the dissolution of high solubility evaporite rocks interspersed with
the Triassic clays. A hydrochemical and temporal characterisation of the effect of evaporites on
the chemical composition of two carbonate aquifers located in Sierra de Grazalema Mountain
Range (Southern Spain) has been undertaken. Results show differences in the intensity of the
evaporite signal as well as in the mineral species responsible for the hydrochemical modifica-
tion: from sulphate-related minerals (gypsum/anhydrite) to rock salt.
11.1 INTRODUCTION
the evaporites contained in this unit leads to a process of cross formational flow,
or karstification processes that resulted in an increase in permeability, at least on a
local basis. The presence of karst features such as dolines, sinkholes, groundwater-
dependent wetlands, pumping wells and saline springs reflect both the karstification
processes and groundwater flow (Carrasco et al., 2005; Snchez et al., 2009).
The presence of evaporite rocks in carbonate sedimentary sequences is a charac-
teristic inherent to the alpine aquifers in southern Spain (Lpez-Chicano et al., 2001;
Moral et al., 2008; Mudarra & Andreo, 2011). Alpine aquifers are generally found
in Jurassic limestones and dolostones overlaying Upper Triassic deposits. Similar geo-
logical configurations can be found in other European regions such as the Northern
Appennines in Italy (Capaccioni et al., 2001), south western Switzerland (Kilchmann
et al., 2004) and Southern France (Aquilina et al., 2002).
The most common mineral species in the Triassic evaporite facies are sulphates
and halides: gypsum (CaSO42H2O), its dehydrated equivalent anhydrite (CaSO4) and
halite or rock salt (NaCl). They all originate from precipitation in over-saturated
water under arid climatic conditions (Goldscheider & Andreo, 2007).
The effect of evaporites on the chemical composition of groundwater stored in
carbonate aquifers is mainly reflected in the concentrations of the ions that consti-
tute the mineral composition of evaporite rocks: SO42-, Ca2+, Na+ and Cl-. This effect
can range from slight increases of some or all of these components to high contents
which determine the chemical facies of the water (Plummer et al., 1990; Capaccioni
et al., 2001; Ma et al., 2011). Since Ca2+ can also come from the dissolution of car-
bonate minerals including calcite, aragonite and dolomite, SO42-, Na+ and Cl- are the
best elements to assess the contribution of evaporite rocks to the final composition
of groundwater draining from carbonate aquifers. Water that was in contact with
evaporites has sometimes also a higher concentration of trace elements (Hunkeler &
Mudry, 2007).
This chapter aims to characterise the effect of the evaporite rocks on the chem-
ical composition of the water drained by karst springs as well as their temporal
evolution and their relationship with the hydrodynamic response. The study area is
the westernmost part of the Sierra de Grazalema Mountain karst system in Southern
Spain.
The area of interest corresponds to the El Bosque and La Silla carbonate aquifers,
which are located in the western sector of the Grazalema Mountain Range in the
province of Cdiz, Southern Spain (Figure11.1).
The relief is rugged with altitudes ranging from less than 300 m asl in the western-
most part of the study area to almost 1000 m asl in the central sector of the El Bosque
aquifer. The latter is separated from La Silla aquifer by a clayey valley (Figure11.1).
The prevailing climate is semi-continental Mediterranean. In general, the annual dis-
tribution of precipitation presents a marked seasonal pattern. The first rainfall nor-
mally takes place in the autumn, at the beginning of the hydrological year. This is
often in the form of intense storms from the Atlantic Ocean, featuring heavy precip-
itation. Winter rainfall and, to a lesser extent, spring rains are commonly associated
The Grazalema Mountain Range (Southern Spain) 157
W E
4.072.000
S SPAIN
sea
ean
ran
ter
Atlantic
M edi
ocean
4.070.000
Grazalema
Mountain
Range
4.068.000
Spring Lithology
4.066.000
with humid winds, which account for most of the annual precipitation. During the
rest of the year, rainfall is scarce to non-existent, particularly in summer. The historic
mean annual precipitation for the period 19652006 ranges from 900 to 1000 mm/
year, with a positive gradient towards the east where values exceeding 1000 mm/year
are recorded. The mean annual temperature in the region is about 16C (Glvez-
Maestre, 2005) and both precipitation and temperature vary with the altitude.
Geologically, the two aquifers consist of rocks belonging to the External Zone
of the Betic Cordillera. The stratigraphic sequence begins with clays and evaporites
from the Upper Triassic age, continues with a thick Jurassic dolostones and limestones
series and finishes with Middle-Upper Jurassic limestones with chert and Lower
CretaceousOligocene marly limestones and marls at the top (Peyre, 1974; Martn-
Algarra, 1987) (Figure11.1). Flysch-type clays and sandstones outcrop in the central
and eastern sector of the area. The structure of both aquifers consists of north west
to south east (El Bosque) and north north west to south south east (La Silla) synclines
the cores of which are occupied by Jurassic-Cretaceous marly-limestones. The entire
folded structure has been affected by more recent fractures, predominantly with north
east to south west and north west to south east directions.
El Bosque and La Silla aquifers are made up of fractured and karstified Jurassic
dolostones and limestones. Adjacent marly-limestones behave as an aquitard.
Impervious lateral boundaries are the Triassic clays (mainly the western and southern
158 Karst without Boundaries
limits), flysch clays to the east and Cretaceous marls (only in El Bosque aquifer). The
base of La Silla aquifer and probably most of El Bosque aquifer consists of Triassic
clays with evaporites.
Recharge to the aquifers takes place exclusively by direct infiltration of rainwa-
ter into dolostones, limestones and, to a lesser extent, limestones with chert, while
discharge is produced through springs located at the edges of the mountains, coin-
ciding with the contact between the permeable carbonate rocks and the aquitards, as
well as by pumping intended for drinking water supply (only in El Bosque aquifer).
Table11.1 summarises the main characteristics of the discharge points that were anal-
ysed. Springs are located at altitudes ranging from 290 to 345 m asl. El Bosque aquifer
has a total area of 18 km2 from which 16 km2 correspond to permeable outcrops,
while La Silla has an area of 11.5 km2, from which 10 km2 are permeable.
Karst surface landforms are not well developed in the area. Karrenfield is rare and
the presence of exokarstic forms such as dolines is scarce. Travertine formations are
recognised in the borders of both aquifers. On the south west edge of El Bosque unit
there is an inactive travertine outcrop indicating the existence of a former drainage
point at this place. The western limit of La Silla aquifer has clusters of small travertine
outcrops, one of them is associated to the discharge of Esparragosilla spring and is
still active.
Vegetation is mostly Mediterranean shrub. No known sources of pollution exist
on the recharge areas apart from some extensive livestock farming.
11.3 METHODOLOGY
Field measurements and spring sampling were undertaken between November 2012
to November 2013. Records were kept of the discharge rate and the chemical com-
position of the water drained by selected springs. At the same time as samples were
taken, water temperature, pH and electrical conductivity were measured in situ, using
portable equipment (WTW Cond 3310, HACH HQ40d), with an accuracy of 0.1C,
0.01 and 1 S/cm respectively. Spring discharge was also measured using a propeller
flowmeter model OTT C31 mostly during low and medium flow conditions, and a
salt-dilution flowmeter model SalinoMADD was used in high flow conditions. On
average, flow measurements and water sampling were performed every 23 weeks,
although during flood events monitoring was conducted more frequently; sampling
was reduced to fortnightly and monthly during low flow periods.
11.4 RESULTS
Table 11.2 Main statistical descriptors of a selection of physico-chemical parameters of all water
samples analysed (period November 2012November 2013). (n) number of samples, (cv)
coefficient of variation, EC (electrical conductivity), T (temperature), Alk (alkalinity).
Fuente Santa n 11 11 11 11 11 11 11 11
Mean 853 17.1 318 72 102 24 121 77
cv (%) 5.1% 1.5% 6.6% 5.5% 8.2% 1.3% 12.1% 9.8%
Quejigo n 24 24 24 24 24 24 24 24
Mean 631 17.1 325 62 96 25 43 30
cv (%) 7.5% 1.1% 3.6% 8.4% 6.5% 2.5% 34.3% 31.9%
Vihuelo n 24 24 24 24 24 24 24 24
Mean 551 17.4 320 57 94 26 17 11
cv (%) 4.2% 1.1% 3.8% 5.4% 9.6% 1.9% 7.7% 4.2%
Mquina n 22 22 22 22 22 22 22 22
Mean 564 17.4 310 89 95 32 11 8
cv (%) 2.3% 2.6% 1.9% 5.2% 4.2% 1.5% 7.4% 2.7%
Esparragosilla n 21 21 21 21 21 21 21 21
Mean 2,243 17.5 278 831 324 68 322 193
cv (%) 5.2% 1.2% 12.4% 12.4% 8.2% 9.1% 6.7% 5.9%
160 Karst without Boundaries
A) 100
Esparragosilla
B) 30
Vihuelo 20 Esparragosilla
Quejigo Fuente Santa
Fuente Santa
3
Mquina Mquina
NO
10 9 Quejigo
1:1
Cl+
Ca
Vihuelo ip
8 sh
+M
4+ on
(Ca2++Mg2+)Alk (meq/l)
7
ti
ela
SO
g
6
icr
etr
5
4 iom
ich
3 Sto
0 0
2
0
10
0
0
10
0
1
0.9
O3
0.8
0.7
HC
Na
SO
Mg
0.6
+K
3+
4
0.5
CO
0.4
0.3
0
0
10
SO42 (meq/l)
100 Ca 0 0 Cl+NO3 100
Figure 11.2
(A) Piper diagram. (B) Binary diagram showing the relationship between SO42 and
non-alkalinity Ca2++Mg2+.
these two components minus the alkalinity, in meq/l. The stoichiometric line 1:1 is
shown in the graph. Water samples from Esparragosilla spring show the highest con-
tents of SO42- and non-alkalinity Ca2++Mg2+, with values exceeding 10 meq/l for both
parameters, whereas the rest of springs display much lower values ranging from 0.5
to 2.5 meq/l for non-alkalinity Ca2++Mg2+ and from 1 to 2 meq/l for SO42-.Water
samples taken at Esparragosilla spring plot on the stoichiometric line 1:1, indicating
that the Ca2+ and Mg2+ are not related to the dissolution of carbonate species (calcite,
dolomite). The increases/decreases in SO42- concentrations registered during the mon-
itoring period are accompanied by similar stoichiometric variations of non-alkalinity
Ca2++Mg2+, suggesting a similar origin for the three components which should be
related to the dissolution of evaporite rocks. The calcium sulphate facies of the water
drained by this spring (Figure11.2A) suggests that the chemical composition is mainly
controlled by the dissolution of sulphate mineral species (gypsum and anhydrite).
Nevertheless, some of the water samples show slight increases in Na+K concentration,
which might be indicative of halite dissolution also contributing to the final chemical
composition of the groundwater drained by Esparragosilla spring.
Water samples from the rest of springs show a different pattern. They do not plot
on the 1:1 stoichiometric line. Most of them are above this line, suggesting an excess
of dissolved Ca2+ and Mg2+ in the water drained by these springs.
In Figure11.3A the SO42-/Cl- ratio (mg/l) has been plotted against Na+ concen-
tration for each water sample analysed, whilst the Cl- concentration is represented
as the diameter of each circle. The line representing the SO42-/Cl- ratio=1 has been
highlighted since it separates samples with higher SO42- (ratio>1) or Cl- (ratio<1) con-
centrations with respect to the other ion. The value of this ratio can be related to the
The Grazalema Mountain Range (Southern Spain) 161
A) 250 B) 600
550 Esparragosilla
225 Esparragosilla
Fuente Santa Fuente Santa
500 Mquina
Mquina
200 Quejigo 450 Quejigo
Vihuelo Vihuelo
175 400
Cl+Na+ (mg/l)
Na+ (mg/l)
125
350 200
28
100
150
75
250 26
100 100
50 50 24
10 50 52 56 60
25
0
0 50 150 700 900
1 3 5 7 9 0 100 200 600 800 1,000
0 2 4 6 8 10
SO42/Cl SO4 (mg/l)
2
Figure 11.3 (A) Diagram showing the SO42-/Cl- ratio versus Na+ concentration and the Cl- content
(circles size). (B) Binary diagram showing SO42- against (Cl-+Na+) concentrations.
Points representing water samples from Mquina spring show a nearly horizontal
trajectory. This pattern is the result of larger SO42- variations throughout the moni-
toring period than those of Cl- and Na+, which remained quite stable (Figure11.3B).
Water samples collected at Esparragosilla and Vihuelo springs have inclined
trajectories. This reflects increases in dissolved SO42- in these springs accompanied
by increases in Cl- and Na+ concentrations, and vice versa. However, the ranges of
variation in both springs are not the same. The difference between the maximum
and minimum values measured in Esparragosilla has been 377 mg/l (SO42-) and
138 mg/l (Cl-+Na+), in contrast to 11.2 mg/l for SO42- and 5.4 mg/l for (Cl-+Na+)
in Vihuelo.
135
Na+ (mg/l)
Cl (mg/l)
90 Na+
200 280 120
80
180 240 105
160 70
450 200 90
Mg2+ 80
25
Ca2+ (mg/l)
400 70 Ca2+
Ca2+ (mg/l)
Mg2+ (mg/l)
24
350 120
60
23
300 50 105
22
250 90
1000 21
Alkalinity (mg/l)
315 SO4 2
400 80
SO42 (mg/l)
Alk
Alkalinity (mg/l)
SO42
SO42 (mg/l)
300 800 Alk 70
350
285 600 60
300
270 400 50
18.0 250
Temperature (C)
T 18.0
EC (S/cm)
2,600
Temperature (C)
EC T
17.5 1,000 17.5
EC (S/cm)
EC
2,400 17.0
17.0 900
2,200 16.5
16.5 800 16.0
500
Discharge (l/s)
Rainfall 100
80
Rainfall (mm)
Discharge (l/s)
400 200
Rainfall (mm)
80 Rainfall
300 Discharge 60 150
60 Discharge
200 40 100
40
100 20 20 50
0 0 0 0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Nov
Dec
2012 2013
2012 2013
(A) Esparragosilla (B) Fuente Santa
C) 60 Cl
75
Na+ 60
Na+ (mg/l)
50
Cl (mg/l)
40 45
30 30
20 15
130 Mg2+ 27
120 Ca2+
Ca2+ (mg/l)
26
Mg2+ (mg/l)
110 25
100 24
90 23
80
375
Alkalinity (mg/l)
SO42 72
Alk
SO42 (mg/l)
350 68
325 64
60
300
56
T
Temperature (C)
800 EC 17.5
EC (S/cm)
700 17.0
16.5
600
16.0
120
Discharge (l/s)
Rainfall (mm)
80 Discharge 150
60 100
40
20 50
0 0
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
2012 2013
(C) Quejigo
Figure 11.4 Temporal evolutions of discharge, water temperature and the principal chemical compo-
nents of the water drained by Esparragosilla (A), Fuente Santa (B) and Quejigo (C) springs.
by Fuente Santa was more mineralised during flooding, when the discharge rate was
maximum. Hence, the highest electrical conductivity values (over 900 S/cm) were
measured in March and April. During the sampling period the mineralisation of water
was mainly controlled by Cl- and Na+ contents; their temporal evolutions and that
of EC are almost similar (Figure11.4B). This dependency is especially evident for the
three first months of sampling, when alkalinity, SO42- and Ca2+ concentrations stayed
almost constant. This indicates that the changes in electrical conductivity observed
during that period (>120 S/cm) were neither caused by the dissolution of carbonate
nor sulphate mineral species.
The hydrograph of Quejigo spring also shows a gradual evolution in which sudden
discharge variations do not occur even after the major rainfall events (Figure11.4C).
Unlike Esparragosilla and Fuente Santa springs, in which recharge coincides with the
drainage of the more mineralised water, at Quejigo spring the recharge tends to dilute
the mineralised water. This dilution, nevertheless, is lagged behind the start of the
rainfall season.
The mineralisation of the water from Quejigo spring is principally controlled
by Cl- and Na+ concentrations. The dilution recorded in January 2013 was pro-
duced by decreased concentrations of both ions, while other chemical components
such as SO42- or Mg2+ showed little variation. In low water conditions the EC
increased and coincided with an increase in both Cl- and Na+ concentrations (their
concentrations doubled in flood conditions) and with a slight increase in SO42-
concentration. This suggests the arrival at the spring of less mineralised water
(mainly due to lower Cl- and Na+ concentrations) during flooding conditions. Thus
recharge promotes the drainage of water from the unsaturated zone of the aquifer
which has not been in contact with the Triassic evaporite rocks located at deeper
zones. The steep decrease in alkalinity during May seems to be the consequence
of a common ion effect caused by a higher rate of gypsum/anhydrite dissolution,
which produces an increase in Ca2+ and SO42- concentrations and the removal of
CaCO3 by precipitation.
11.5 DISCUSSION
The mean annual temperature of the water drained by all springs was stable during
the monitoring period (coefficients of variation ranging from 1.1 to 2.6%, Table11.2)
and slightly higher than the average air temperature at the study area (16C). These
increased spring water temperatures suggest relatively deep groundwater flow through
the saturated zone of the aquifers, where water temperatures are higher than in the
shallower recharge parts of the system.
Groundwater stored in the aquifer drained by Esparragosilla spring is in direct
contact with the Upper Triassic deposits which constitute the lower and lateral hydro-
geological limits of La Silla aquifer (Figure11.1). Thus, dissolution of evaporites
mainly gypsum/anhydrite and halite occurs and, as a consequence, water becomes
more mineralised principally because of large increases in the concentrations of
SO42-, Ca2+, Na+ and Cl-. The water from this spring is the most mineralised and
has the greatest SO42-, Na+ and Cl- concentrations (Table11.2, Figure11.3A). This
suggests the existence of groundwater flowpaths through the more permeable parts
The Grazalema Mountain Range (Southern Spain) 165
of the Upper Triassic body, or longer residence times of groundwater in contact with
the evaporites.
The hydrochemistry of Esparragosilla spring water is influenced by both gypsum
and halite dissolution (Figure11.3B), although their contribution is not similar. Mean
SO42 concentrations are considerably higher than those of Cl- and Na+ (Table11.2),
and all samples are of a calcium sulphate facies (Figure11.2A). These data indicate
that the chemical composition of the groundwater drained by this spring is principally
controlled by the dissolution of gypsum/anhydrite.
The temporal evolution of the chemical composition of Esparragosilla spring
water is mainly controlled by a piston-flow effect in the saturated zone of the aquifer
(Figure11.4A). Recharge produces the mobilisation of water from deeper zones of the
aquifer where groundwater is characterised by longer residence times in contact with
the evaporites and higher mineralisation, temperature and SO42-, Ca2+ and Mg2+ con-
centrations. The evaporites located at the bottom of the aquifer drained by this spring
contain both gypsum and halite. While the effect of the dissolution of these minerals
on the hydrochemistry of the water is evident, the role played by carbonate-related
minerals in the temporal evolution of the water chemistry of this spring is much more
attenuated. The hydrodynamic and hydrochemical evolution of Esparragosilla is
characteristic of aquifers with a low degree of karstification whose output signal (flow
discharge) tends to smooth the input signal (rainfall).
The effect of the Triassic substratum on the chemical composition of the water
drained by Fuente Santa and Quejigo springs (El Bosque aquifer) is mainly the result
of dissolution of halite. The water samples collected at these discharge points show the
highest Cl- and Na+ contents with respect to SO42- concentrations (Figure11.3A). In
the case of Fuente Santa the mean concentrations of Cl- and Na+ during the monitor-
ing period have exceeded that of SO42- (SO42-/Cl- ratio <1 in Figure11.3A). The con-
tribution of halite to the hydrochemistry of Quejigo spring water is less evident since
Cl and Na+ concentrations are lower and the SO42-/Cl- ratio is higher (Figure11.3A).
In both springs the annual variations of Cl- and Na+ concentrations are larger than
that of SO42- (Figure11.3B), suggesting a limited and not variable contribution of the
latter (i.e. gypsum/anhydrite) to the final chemical composition of the water flowing
out of these two springs.
In Fuente Santa spring recharge produces the drainage of more mineralised water
stored in deeper zones of the aquifer, where it has been in contact with Triassic evap-
orites for long periods (Figure11.4B). The relatively high mean temperature of the
water (17.1C) also suggests the existence of groundwater flow paths through deeper
parts of the aquifer.
Recharge produces the dilution of the water drained by Quejigo spring
(Figure11.4C). This is consequence of the mixing of groundwater stored in the
saturated zone of the aquifer with more recent and less mineralised infiltrating
water coming from the unsaturated zone of the aquifer. The chemical composition
of the recently infiltrated water is characterised by lower Cl- and Na+ concentra-
tions (Figure11.4C). The EC of the spring water is principally dependent upon the
concentrations of these two components. The final chemical composition of the
water is controlled by the dissolution of halite as well as by other hydrochemical
processes such as the common ion effect, which can provoke relatively sudden
reductions in alkalinity.
166 Karst without Boundaries
11.6 CONCLUSIONS
ACKNOWLEDGEMENTS
REFERENCES
Aquilina L., Ladouche B., Doerfliger N., Seidel J.L., Bakalowicz M., Dupuy C., Le Strat P. (2002)
Origin, evolution and residence time of saline thermal fluids (Balaruc springs, southern France):
implications for fluid transfer across the continental shelf. Chemical Geology 192, 121.
Barber J.A., Andreo B., Almeida C. (2014) Using non-conservative tracers to characterize
karstification processes in the Merinos-Colorado-Carrasco carbonate aquifer system (south-
ern Spain). Environmental Earth Science 71, 585599.
Capaccioni B., Didero M., Paletta C., Salvadori P. (2001) Hydrogeochemistry of groundwaters
from carbonate formations with basal gypsiferous layers: an example from the Mt Catria Mt
Nerone ridge (Northern Apennines, Italy). Journal of Hydrology 253, 1426.
Carrasco F., Andreo B., Linares L., Snchez D., Rendn M., Cobos A., Ortega F., Vadillo I.,
Prez I. (2005) Contexto hidrogeolgico de humedales del norte de la provincia de Mlaga
[Hydrogeological context of wetlands located in northern Mlaga province]. In: Lpez-Geta
J.A., Rubio-Campos J.C., Martn-Machuca M. (editors) VI Simposio del Agua en Andaluca,
Sevilla, Espaa. Serie Hidrogeologa y Aguas Subterrneas, vol. 14. Publicaciones del IGME,
605618.
Glvez-Maestre M.J. (2005) Climatologa [Climatology]. In: Lpez-Geta J.A. (editor) Atlas
Hidrogeolgico de la Provincia de Cdiz. Instituto Geolgico y Minero de Espaa y
Diputacin de Cdiz. Madrid, Spain, 5358.
Goldscheider N., Andreo B. (2007) The geological and geomorphological framework. In:
Goldscheider N., Drew D.P. (editors) Methods in Karst Hydrogeology. Taylor & Francis,
London, United Kingdom, 924.
Hunkeler D., Mudry J. (2007) Hydrochemical methods. In: Goldscheider N., Drew D.P. (edi-
tors) Methods in Karst Hydrogeology. Taylor & Francis, London, United Kingdom, 93121.
Kilchmann S., Waber H.N., Parriaux A., Bensimon M. (2004) Natural tracers in recent ground-
waters from different Alpine aquifers. Hydrogeology Journal 12, 643661.
Lpez-Chicano M., Bouamama B., Vallejos A., Pulido-Bosch A. (2001) Factors which deter-
mine the hydrochemical behaviour of karstic springs: a case study from the Betic Cordilleras,
Spain. Applied Geochemistry 16, 11791192.
Ma R., Wang Y., Sun Z., Zheng C., Ma T., Prommer H. (2011) Geochemical evolution of
groundwater in carbonate aquifers in Taiyuan, northern China. Applied Geochemistry 26,
884897.
Martn-Algarra M. (1987) Evolucin geolgica alpina del contacto entre las Zonas Internas
y Externas de la Cordillera Btica [Alpine geological evolution of the contact between the
Internal and External Zones of the Betic Cordillera]. PhD thesis, University of Granada.
1171 p.
Martn-Algarra M., Vera J.A. (2004) Evolucin de la Cordillera Btica [Evolution of the Betic
Cordillera]. In: Vera J.A. (editor) Geologa de Espaa. Sociedad Geolgica de Espaa e
Instituto Geolgico y Minero de Espaa, 437444.
Moral F., Cruz-Sanjulin J., Olas M. (2008) Geochemical evolution of groundwater in the car-
bonate aquifers of Sierra de Segura (Betic Cordillera, southern Spain). Journal of Hydrology
360, 281296.
168 Karst without Boundaries
Mudarra M., Andreo B. (2011) Relative importance of the saturated and the unsaturated zones
in the hydrogeological functioning of karst aquifers: the case of Alta Cadena (Southern
Spain). Journal of Hydrology 397, 263280.
Peyre Y. (1974) Gologie dAntequera et de sa rgion (Cordillres Btiques, Espagne) [Geology
of Antequera and its region (Betic Cordillera, Spain)]. PhD thesis, Institute National
Agronomique, Paris. 528 p.
Plummer L.N., Busby J., Lee R., Hanshaw B. (1990) Geochemical modelling of the Madison
aquifer in parts of Montana, Wyoming, and south Dakota. Water Resources Research 26,
19812014.
Snchez D., Carrasco F., Andreo B. (2009) Proposed methodology to delineate bodies of
groundwater according to the European water framework directive. Application in a pilot
Mediterranean river basin (Mlaga, Spain). Journal of Environmental Management 90,
15231533.
Chapter 12
ABSTRACT
It is very difficult to define hydrogeological parameters in an anisotropic aquifer such as
carbonate matrix. Even the application of many different methods cannot provide a robust
case for the karst aquifer. However, in the past decade a whole new branch of karst investiga-
tion, 3D modeling of karst aquifers and karst conduits, has become widely used in the earth
sciences. The main topic of this chapter is 3D reconstruction and modelling of the Beljanica
karst aquifer. The input parameters of the model are represented by the 69 caves, 15 sink-
holes and 1682 dolines, 7 major karst springs, around 70 minor springs, and all available
geological, morphological and hydrogeological data. As output from the model, more than
6000 nodes (registered in the database) were calculated. There are in some cases 60 different
numbers for each node. The 3D model of the karst conduit network, enables the reconstruc-
tion of a complex karst aquifer, including its spatial, temporal, quantitative and qualitative
characteristics.
12.1 INTRODUCTION
sections, caves, sinkholes, springs and karst channels were converted into digital
form, and each spatial unit was defined by its x, y and z coordinates. The compila-
tion of all elements in the 3D environment produced a real, spatially oriented net-
work of potential karst-conduit pathways. The case study of the Beljanica aquifer
generally shows the procedure and results of the 3D modelling methodology of karst
conduits, and further analysis for the purpose of watershed delineation, water path-
way determination and assessment of karst reservoir storativity and consequently
static and dynamic reserves.
Analysing the geometry of the main karst conduits in the saturated zone and the
connection with conduits in the aeration zone, provides insight into karst channels
and has enabled the creation of a 3D model. Analysis of various parameters obtained
by quantitative and qualitative monitoring of groundwater characteristics and their
analysis through the established physical model provides data on the relationship
between the water recharge and discharge zones. Such a model can be further used for
an analysis of the speleogenesis and hydrogeology of the area.
Generally there are three zones of research in karst aquifer spatial modelling.
The first is the recharge zone and all underground forms that are related to this
area such as dolines, sinkholes, caves and pits. The third zone is the phreatic zone
with main forms such as karst springs and their submerged karst channels that
often lead to the deeper parts of karst aquifers. The most problematic is the sec-
ond zone between these two sections, but with the support of different software
for 3D modelling and knowledge of karst development and intensity, this part can
be reconstructed and the pathways (conduits) connections from inlet to outlet can
be made.
The basic problem in the determination of the methodology was how to perform
a quality analysis, or how to state the problem whose goal was to develop a model
which would be an analysis of the geometry of karst conduits, integrating hydrogeo-
logical laws and geological characteristics.
The methodology is illustrated with the case study of the Beljanica karst aquifer
(area of around 300 km2). This study shows the procedure and results of the 3D
modelling methodology for karst conduits, and further analysis is made for the pur-
pose of watershed delineation, water pathways and the definition of karst reservoir
storativity.
Data entry is the starting point for the establishment of a karst aquifer model and
includes the conversion of available data into a digital format (graphic and alpha
numeric). Such a model-designed system includes four main elements:
Main layers collected (as raster and vector entities) from detailed field studies are:
1. Topographic layer
1. DEM (Digital Elevation Model)
1.
DEM of Surface
2.
DEM of Base of Karstification
2. Geological map
3. Hydrogeological map
4. Map of sinkholes and dolines
5. Map of groundwater levels
6. Map of forests, small plants and bare karst
7. Map of cave channels distribution
8. Tectonic map
9. Map of base of karstification
10. Map of the hypsometric location of the sinkholes, caves and karst springs.
The model with all the layers can generally be shown through the relationship of the
spatial entities as a single model (Figure12.1).
Components of the model include all of the classes and entities which, through a series
of computational network and geostatistical analyses, form the 3D output of the Beljanica
karst aquifer. The result of the output data is the connection of surface spatial-oriented
data with the defined position of karst channels in the karst systems that are in connection
between the implemented database and the spatial 2D and 3D entities (Figure12.1).
Network spatial data 2D and 3D entities represent their field of numerical clusters
that are 2D data transformed into a 3D object through a simulation model. 3D objects
belong to the multipatch feature class and contain three-dimensional information.
Dreinage network
DEM
Caves, pits, sinkholes
3D channels distibution
Hydrogeological objects
(boreholes wells)
Hydrogeological features
(spring, sinkholes)
Karst aquifer
volume
Tectonic
GWL
Base of karstification
The Beljanica karst massif is located in the eastern part of Serbia (Figure12.3), and
the mountain range is an anticline (Kucaj-Beljanica structure) composed of Jurassic
and Cretaceous carbonate rocks generally inclined from the central to peripheral parts
of the massif (Stevanovic, 1991). The carbonate rock complex of the Beljanica mas-
sif is the result of the Alpine orogenic phase. Tectonic events resulted in a complex
system of faults and fractures which are preferred subterranean water flow path-
ways. The main forms are dolines, caves, pits, sinkholes and karstic springs and their
3D Spatial modelling of karst channels The Beljanica karst massif 173
+
+
+
+
2D map of observed and Creating 3D
+ calculated model points conduit network.
+
+
KARST
SPRING
Figure 12.2 Schematic procedure of creating the 3D physical model (Milanovic, 2015).
classification and exploration was of great importance in the process of creating the
Beljanica karst aquifer physical model. The main hydrogeological characteristic of the
aquifer is its deep circulation. Cave divers explored channels from the Krupaja Spring
to a depth of 133 m and Mlava Spring to a depth of over 73 m (Figure12.4), the latter
channels go deeper (survey not yet completed) (Milanovic, 2007).
This case study includes an analysis of the 3D ArcGIS physical model of the
karst system. The input data for the model are represented by the 69 caves, 15 sink-
holes and 1682 dolines, 7 major karst springs and around 70 minor springs that were
also explored plus the geological, morphological and hydrogeological data connected
with the area of investigation (Figure12.5A). As 2D output from the model, more
than 6000 data points (registered in the database) were calculated. There are 60 dif-
ferent data for each point (coordinates, groundwater level, type of channel, dimen-
sions of channels, next point of conduit, connection with (conduit), orientation of
channel, hydrogeological function etc.). Eventually, the model is determined by more
than 360 000 data inputs (Milanovic et al., 2010). The 2D model mesh is shown on
Figure12.5B.
The total length of the karst channel network, which is calculated using this
model and presented in a 3D environment, is 647.3 km. Detail for the model are
shown on Figure12.6.
7400000000000 7500000000000 7600000000000
5100000000000
5100000000000
DEM Serbia
Legend
Value
High : 2600 m a.s.l.
Low : 50 m a.s
5000000000000
5000000000000
4900000000000
4900000000000
4800000000000
4800000000000
4700000000000
4700000000000
7400000000000 7500000000000 7600000000000
Legend
Karst Spring Karst
Small Spring Beljanica
Ponor
GW direction Cave
River
Figure 12.3 Geographic position of Beljanica massif with main ponors and karstic springs.
3D Spatial modelling of karst channels The Beljanica karst massif 175
5m
ENTRANCE 0m
10
20
30 m entrance to 30
siphonal channel
40
50
Aluvial deposit
Neogen deposit Spring
60
Fault
? 70
73 m depth of
Assum direction of investigated channel ? 75
main karstic channel
Figure 12.4 Result (cross section) of Mlava spring cave diving investigation.
A) B)
Figure 12.5 Map of Beljanica karst massif including position of hydrogeological, geological, geomorpho-
logical and speleological features A, 2D model mesh according Beljanica physical model
data B.
A spatial-oriented network, the potential karst channels providing a base for the
assessment of storativity, is a crucial component for groundwater resources analysis.
The underground karst development along with the saturation status were important
elements from the 3D model to be correlated with data collected from simultaneous
groundwater quantity (discharge) observations (Milanovic et al., 2013).
176 Karst without Boundaries
GWL max
GWL min
Figure 12.6 3D model of conduit system of the Beljanica karst aquifer with Mlava spring physical
model.
Figure 12.7 Delineation of watershed of Beljanica karst aquifer through analysis of karst channels
distribution (3D analyses).
REFERENCES
Borghi A., Renard P., Jenni S. (2010) How to model realistic 3D karst reservoirs using a
pseudo-genetic methodology Example of two case studies, Advance in Research in Karst
Media, Springer.
Butscher C., Huggenberger P. (2007) Implications for karst hydrology from 3D geological
modeling using the aquifer base gradient approach, Journal of Hydrology 342, 184198.
Fish L. 1996. Compass, http://members.iex.net/lfish/compass.html.
Filipponi M., Jeannin P.Y. (2008) Possibilities and Limits to predict the 3D Geometry of Karst
Systems within the Inception Horizon Hypothesis, Geophysical Research Abstracts, Vol. 10,
EGU General Assembly 2008, EGU2008-A-02825.
Filipponi M. (2009) Spatial Analysis of Karst Conduit Networks and Determination of
Parameters Controlling the Speleogenesis along Preferential Lithostratigraphic, Horizons,
thse no 4376, Suiss.
Gogu R. C., Carabin G., Hallet V., Peters V., Dassargues A. (2001) GIS-based hydrogeological
databases and groundwater modelling, Hydrogeology Journal 9, 555569.
Jeannin P.Y., Groves C., Hauselmann P. (2007) Speleological investigations, Methods in karst
hydrogeology, IAH Book Series, Taylor & Francis, London.
Kincaid T.R. (2006) A method for producing 3-D geometric and parameter models of satu-
rated cave systems with a discussion of applications, Groundwater Flow and Contaminant
Transport in Carbonate Aquifer, Taylor & Francis, 2000, pp. 169190.
Kovacs A. (2003) Geometry and hydraulic parameters of karst aquifers: a hydrodynamic
modeling approach, Doctoral theses, University in Neuchatel.
Milanovic S. (2007) Hydrogeological characteristics of some deep siphonal springs in Serbia
and Montenegro karst. Environmental Geology. 51(5), 755759.
178 Karst without Boundaries
Milanovic S. (2010) Creation of physical model of karstic aquifer on example of Beljanica Mt.
(eastern Serbia), Doc. dissert, FMG, University of Belgrade, Belgrade.
Milanovic S., Stevanovic Z., Vasic Lj. (2010) Development of karst system model as a result of
Beljanica aquifer monitoring. Vodoprivreda Vol. 0350-0519, 42 (2010) pp. 209222.
Milanovic S., Stevanovic Z., Vasic Lj., Ristic-Vakanjac V. (2013) 3D Modeling and monitoring
of karst system as a base for its evaluation and utilization A case study from eastern Serbia,
Environmental Earth Science, 71(2), 525532.
Milanovic S. (2015) Physical modeling of karst environment karst aquifers. In: Karst Aquifers
Characterisation and Engineering. (Ed. Z. Stevanovic), Springer, Professional Practice in
Earth Sciences, pp.267281.
Ohms R., Reece M. (2002) Using gis to manage two large cave systems, wind and jewel caves,
south dakota, Journal of Cave and Karst Studies, April 2002.
Stevanovic Z. (1991) Hydrogeology of Carpathian-Balkan karst of eastern Serbia and water
supply opportunities (in Serbian). Spec. ed. Fac. Min. & Geol., pp. 1245, Belgrade.
Strassberg G. (2005) A geographic data model for groundwater systems, Doctoral thesis, The
University of Texas at Austin December, 2005.
Chapter 13
ABSTRACT
Hydrogeological and geological field surveys are the indispensable basis for purposeful mon-
itoring campaigns and catchment characterisation. In order to acquire, within a given time-
frame, comprehensive information about complex hydrogeological settings or under-studied
catchments it is not only advisable, but essential to follow a step-by-step protocol that is based
on goal-oriented field work. Information from such an approach applied to the delineation
of catchment characteristics enables the conduct of process-oriented monitoring at strategic
locations and a more meaningful interpretation of resulting data. The approach described in
this chapter becomes particularly important regarding the complexity of karstified aquifers
and overlying catchments. Implementation of the methodology was applied to investigations
in a karst area within the Outer Dinarides (Gorski Kotar, Western Croatia). In this study,
information about the complex local hydrogeology was obtained through geological mapping,
hydrogeological field surveys, and hydrological monitoring using stableisotopes.
13.1 INTRODUCTION
13.1.1 Objectives
The aim of the study is to describe, how comprehensive information about the
complex karst system around Zeleni Vir spring was gained, particularly regarding
discharge and storage dynamics, as well as catchment- and aquifer characteris-
tics. These were obtained by applying a process-orientated scheme of combined
hydrogeological field methods, such as mapping springs, hydrogeological quality
rating of rocks, geological investigations on local lithologies, and stable isotope
hydrology.
180 Karst without Boundaries
spring
stream
water
fall
cave
sampling
point
Figure13.1 South view into the Curak valley. Map showing lithology, springs, streams and sampling
locations.
A field work oriented approach for complex karst aquifer characterisation 181
13.2 METHODS
The methodological concept follows the scheme of prior assessing local geology and
hydrogeology by field surveys, to gain an overview about the catchment and then
to determine the most useful locations for monitoring and sampling (Figure13.2).
This approach will support an emerging insight to the complex hydrogeology of the
research area, allow a focused data interpretation and consolidate results from field
survey and lab analyses in a comprehensive way.
13.2.2 Hydrogeology
A hydrogeological quality rating of rocks in the catchment was conducted to obtain
an overview of the hydrogeological characteristics of local lithologies. This method is
focused on the capability of rocks to act as a potential aquifer or aquiclude. Therefore,
peculiarities of rocks such as grade of karstification, open joints and infiltration capac-
ity are important, rather than petrographic or stratigraphic aspects. Seven represen-
tative outcrops in the catchment (Triassic) and spring (Jurassic) area were assessed
(Figure13.3). Bedding thickness, joint frequency, number of water-transmitting joints
and bedding planes, unfilled joints and karstification grade were determined for each
outcrop every 5 m.
Geological and hydrogeological information, including discharge (Q), electrical
conductivity (EC) and temperature, was used to identify springs and streams of inter-
est for subsequent monitoring and sampling.
Jurassic
Triassic
Figure13.3 Geological map (modified after Savic & Dozet, 1985) of the research area with recorded
outcrops, showing the Zeleni Vir window and indicating the Karst water flow under-
neath the confining Permian formation. Dashed line marks the cross section shown in
Figure13.10.
assumed to be a karstic aquifer, discharging at the contact between Karst floor 2 and
Permian formations. During the monitored precipitation event the entire discharge of
Zeleni Vir spring was captured in the power plants pressure pipeline, the overflow
was not active and, therefore, only water originating from the springs close to Skrad
was monitored at the sampling location Zeleni Vir stream. The second sampling
location was chosen at Devils Creek (DC), which is dominated by surface runoff
from a nearby gorge, discharging from the capping Permian clastic rocks. Some 50m
downstream of the sampling locations both streams flow together, forming the Curak
river (Figure13.1).
Samples were taken manually with an interval from one hour to three hours,
covering the whole dynamics of the event. Water temperature and conductivity were
measured in hourly intervals. The precipitation event consisted of a short-duration
thunderstorm followed by steady rain. Rainfall data were collected continuously for
the duration of the event (48 h), during which a total 29 mm of rainfall was measured.
A gauge was installed at the Devils Creek sampling location and runoff was
measured applying the tracer dilution method (Hubbard et al., 1982; Benischke &
Harum, 1984) during critical moments of the event. Following this, a discharge stage
relation was calculated using six measurements.
A second gauge was installed at the Curak river where runoff was calculated
applying the Manning-Strickler formula as described in Strobl & Zunic (2006).
The sampling campaign started in February 2011 and lasted nine month. Thirty
water samples of Zeleni Vir spring were taken on a regular basis and four precipita-
tion samples during diverse seasons were gathered.
The isotopic composition of the water samples was measured in the labora-
tory by using cavity ring-down spectroscopy (Berden et al., 2001) with a WS-CRDS
(Wavelength-Scanned Cavity Ring-Down Spectroscopy) instrument of Picarro, Inc.
coupled to a CTC HTC-Pal liquid autosampler (LEAP Technologies, Carrboro, NC,
USA) for automated measurements of liquid water samples. CRDS is a direct absorp-
tion technique (Berden et al., 2001) that offers results for pure water samples highly
comparable in precision with classical mass spectroscopy (Brand et al., 2009). The
instrumental setup is comparable to the system described by Gupta et al. (2009).
13.3 RESULTS
Jurassic
Triassic
Permian
Figure13.4 Stereo plots showing bedding- (1st plot each) and fault planes (2nd plot each).
A field work oriented approach for complex karst aquifer characterisation 185
K2: 030/90). Caves, like the Muzeva Hiza cave and the Zeleni Vir Cave are linked
to strained and notably karstified zones and bound to the major fold systems of the
Gorski Kotar.
Triassic, Outcrop No.: 9, 10, 19, 20, 21 (Figure13.3):
Triassic carbonates are mainly north west to south east striking and north to
north east dipping (Figure13.4). At outcrops, distinct, partly conjugated fault systems
can be found. Slickensides and fracture traces are present. Karstification is strongly
associated with tectonically strained zones, but also cleavages are considerably karst-
ified. The fractured Triassic limestones around Ravna Gora are, due to the high den-
sity of faults and joints with significant karstification, a catchment area with high
infiltration capacity.
Permian, Outcrop No.: 12, 13, 14 (Figure13.3):
Outcrops in the Permian formation are rare. Nevertheless the distribution of
Paleozoic clastic rocks can be tracked because of its smooth morphology that dif-
fers explicit from that of karst. Bedding planes dip around 30 north (Figure13.4).
The tectonic contact between the overthrusted and confining Permian rocks with the
Jurassic carbonates of the footwall can be observed at an outcrop near Dolnje Tihovo
(outcrop No. 14, Figure13.3).
Figure13.5 Example pictures of the five types of hydrogeological rock quality, determined in the
research area.
Type 1: Limestone with intact bedding, bedding-planes are unfilled and karstified.
This potential aquifer is characterized by intact limestone with a bedding thick-
ness of several meters. The open and unfilled bedding planes are characterised by
slight dipping and show distinct signs of karstification. Due to intact and gently dip-
ping beddings, Type 1 rocks are assumed to have minimal potential for direct infiltra-
tion of precipitation water and act predominantly as an aquifer allowing lateral water
flow. It can be assumed the aquifer of Zeleni Vir spring exhibits similar properties, at
least in part.
Type 2: Highly fissured carbonates with discordant thrust faults and water-transmitting
joints.
Highly fissured and strained rocks with water-transmitting joints and discordant
fault systems. These rocks allow an interchanging flow between different karst water
levels, due to discordant faults.
Type 3: Fissured carbonates with closed or filled joints.
Dolomite rocks and limestones with closed joints are assumed to act in a manner
similar that of aquiculdes. Much of the flow occurs in the Type 1 and Type 2 rocks.
Type 4: Fissured limestones and highly developed karstification.
A field work oriented approach for complex karst aquifer characterisation 187
The genesis of spelean formations; cavities and caves are mainly linked to these
highly fissured and karstified limestones. Type 4-rocks can be found in the spring area
e.g. in the surrounds of the cave Muzeva Hiza, located in the upper part of Devils
Creek gorge. In the catchment area, Type 4 rocks are assumed to enable dominant
parts of infiltration. Filling of joints and cavities are often palaeosols that contain
iron-rich earth. The geomorphology typically involves karstification characterised by
frequent occurrences of dolines and ponors.
Type 5: Highly fissured Dolomite with closed or filled joints.
The appearance of dolomite is limited to small scaled areas. The infiltration
capability is assumed to be low and the associated landscape does not show the asper-
ity that often characterises karstified limestones.
CG - C B
Qi = QG (13.2)
Ci - CB
Table13.2 Environmental isotope- and hydrologic data from event monitoring at sampling location
Zeleni Vir Stream.
Water
Date Sample ID EC [S/cm] temp. [C] Q [l/s] O18 D d
For Devils Creek precipitation water (Qi) contributed a maximum of 21% to total
discharge, whereas in the Zeleni Vir stream a maximum of 9% precipitation water
could be detected (Figure13.8). The maximum event water participation in the Devils
Creek was determined 1.5 hours later than in the Zeleni Vir stream (Figure13.8).
Following the two components isotopic separation method the interflow was cal-
culated from the difference between total discharge and event water in the stream.
This again shows clearly a time shift between the different components. At the Devils
Creek sampling location the interflow reached its peak 3 hours before the event water
component, showing a time shift between hydraulic reaction and maximum rainwater
amount of the spring system in the Permian aquifer (Figure13.9). After the surface
flow passage in the Devils Creek, this signal is dispersed.
Analyses of water samples reveal distinct differences in hydrochemical parame-
ters of both monitored streams (Table13.5). Water from Devils Creek (Table13.5,
A field work oriented approach for complex karst aquifer characterisation 189
Table13.3 Environmental isotope- and hydrologic data from event monitoring at sampling location
Devils Creek.
DEVIL'S CREEK
Water
Date Sample ID EC [uS/cm] temp. [C] Q [l/s] O18 D d
DC), discharging the Permian clastics has a lower mineral content (A-sum, K-sum)
than does the Zeleni Vir stream (Table13.5, ZV), discharging an carbonatic aquifer.
This validates the intended selection of sampling locations and also confirms on-site
electrical conductivity measurements during baseflow conditions for both streams
(DC:200S/cm, ZV:307S/cm). The Mg, Ca and HCO3 concentrations and espe-
cially the Ca/Mg ratio of 3.4 for Zeleni Vir stream water determines its hydrochemical
facies to carbonate rocks, most likely Dolomite (Pavuza & Traindl, 1983).
9.50
18O%
10.00
10.50
15.06 12:00 16.06 04:00 16.06 20:00 17.06 12:00
Figure13.6 18O values of Devils Creek and Zeleni Vir stream during the event monitoring.
signal was determined for samples taken in September 2011 (Table13.4). The deu-
terium excess ranges from 14.0 in July 2011 to 17.1 in February 2011, with a mean
value of 15.7. Complementary to the spring water, rain water was sampled (n = 4) in
Zeleni Vir. The precipitation showed high 18O values during the early summer month
(samples from June and July 2011), when a signal of -5.7 to -5.3 was measured.
Significantly depleted 18O values of -11.5 were determined from samples taken
in September 2011. The deuterium excess in precipitation water ranges from 12.0 in
June 2011 to 17.2 in September 2011, with a mean value of 14.
Although no precipitation samples from the catchment area of Zeleni Vir spring
were available, a Local Meteoric Water Line (LMWL) can be estimated as follows:
assuming that the spring water of Zeleni Vir is a mixture of precipitation infiltrated,
without major fractionation in the hydrogeological catchment around Ravna Gora
and the measured values of the isotope ratio are scattered around a LMWL for Ravna
Gora, a substituted LMWL can be derived:
2H=7.918O+13.0 (13.3)
Evaluation of isotope data from spring water samples taken at Zeleni Vir spring
revealed a range of 18O values significantly lower (B = 0.8 ) than those determined
for monitored summer precipitation (A = 6.3 ). This may reflect a high reservoir
capacity of the aquifer and low dynamics. Partly this may be caused by the relatively
small amounts of summer precipitation compared with the annual mean. Accordingly
the recharge of Zeleni Vir spring is mainly driven by precipitation during the autumn
and winter months. As described by Clark & Fritz (1997) the strong correlation
A field work oriented approach for complex karst aquifer characterisation 191
Table13.4 Environmental isotope data from long term sampling campaign at Zeleni Vir Spring.
m = c -1 f -2 - 1 (13.4)
Where m is the calculated mean transit time in days, c is the angular frequency
constant (2p /365) and f is the damping coefficient (f = B/A).
Although a higher number of spring water and precipitation samples are prefera-
ble, this analytical solution still enables an estimate of the mean transit time. Making
use of the available data, a MTT for Zeleni Vir spring water of 1.2 years can be
calculated.
18O%
10.50 10.00 9.50 9.00
62.0
2H%
67.0
72.0
Devils Creek Zeleni Vir Stream Zeleni Vir Spring LMWL (Mt. Velebit) LMWL (Ravna Gora)
Figure13.7 Isotope ratio of deuterium and 18O at both sampling locations during the event monitor-
ing, isotope ratio of Zeleni Vir spring (long termed sampling campaign), LMWL calculated
for Ravna Gora and LMWL for Mount Velebit (Vreca et al., 2006).
15.00
Qi %
10.00
5.00
Figure13.8 Percentage of event water contributing to the discharge of both streams during the event
monitoring.
A field work oriented approach for complex karst aquifer characterisation 193
200.0
interflow peak
07:15
150.0
Q [l/s]
100.0
0.0
15.06 12:00 16.06 00:00 16.06 12:00 17.06 00:00 17.06 12:00 18.06 00:00
Figure13.9 Component separation for the Devils Creek showing a distinct time shift between inter-
flow and event water.
The study followed the proposed approach based on selected field methods, suitableto
gather information about the complex hydrological and hydrogeological setting of the
Zeleni Vir karst area (Figures13.2 and 13.3). Geological mapping provided a funda-
mental understanding of local stratigraphy and tectonics (Figures13.1 and 13.10).
Hydrogeological investigations, such as the hydrogeological quality rating of rocks
enabled to determine their specific characteristics, governing their overall function
as either an aquifer or an aquiclude (Table13.1, Figure13.5). Mapping of springs
completed this task. This target-oriented survey led to an overview of the local karst
hydrology and enabled the selection of strategic sampling locations for subsequent
monitoring programmes (Figure13.1). Due to the merging of results that were based
on field work with those from laboratory analyses consolidating information about
(i) Zeleni Vir spring, (ii) Devils Creek and (iii) Zeleni Vir stream was gained, provided
information on the local karst hydrogeology.
(i) The estimated LMWL for Ravna Gora lies within the range of the LMWL cal-
culated for Mount Velebit (Vreca et al., 2006) (Figure13.7). This confirms the
hypothesis that the precipitation in coastal mountain ranges of the Dinarides, like
Gorski Kotar and Mount.Velebit emerges mainly from air masses having a com-
parable origin. Enriched isotopic signatures for rain water at Zeleni Vir during
the summer months can be ascribed to precipitation that originates from Adriatic
air masses, whereas decisively lower enrichments are caused by the increased
Table13.5 Results from hydrochemical analyses of samples taken at Devils Creek and Zeleni Vir stream during the event monitoring.
DC4 16.06.2010 06:30 4.72 0.94 6.98 19.41 4.15 1.05 8.82 92.1 1.77 1.83 -0.06 3.11
DC12 16.06.2010 23:00 5.29 0.95 7.61 21.00 4.76 1.00 9.10 99.5 1.93 1.97 -0.04 2.12
DC16 17.06.2010 16:00 5.44 0.88 7.98 21.78 4.87 0.98 9.41 104.3 2.00 2.06 -0.06 2.80
ZV4 16.06.2010 06:30 3.14 0.42 12.59 42.75 4.77 2.74 6.25 183.7 3.32 3.32 0.00 0.09
ZV12 16.06.2010 23:00 3.28 0.50 13.00 43.34 4.78 2.83 6.51 186.7 3.39 3.38 0.01 -0.32
ZV16 17.06.2010 16:00 3.24 0.63 12.96 43.33 4.90 2.81 6.61 186.1 3.38 3.37 0.01 -0.40
A field work oriented approach for complex karst aquifer characterisation 195
Figure13.10 Schematic stratigraphy of the Zeleni Vir window. Compare cross section with Figure13.3.
discharging from a fissured, but not karstified aquifer (Type 5-rock: Figure13.5,
Table13.1). The results of hydrochemical analysis indicate a carbonate aquifer,
most likely dolomitic. 18O values for the sampled water were used to estimate
the height of the catchment. As the altitude of the Devils Creek orographic catch-
ment can be assigned to confine the Permian formation and 18O depletion varies
between 0.15 and 0.5 per 100 m rise (Clark & Fritz, 1997), the catchment
location for Zeleni Vir stream was estimated using 18O data and found to be at
the altitude were the contact of Triassic formations (Karst floor 2) and Permian
rocks occurs (Figure13.1). Further investigations would be desirable to gain more
accurate information about location and dimensions of this aquifer.
The investigations show again, like other studies conducted in the Dinaric karst
(e.g. Herak, 1980; Biondic et al., 1997; Kresic & Stevanovic, 2010), that structural geo-
logical settings are affecting the karst hydrology of western Croatia on diverse scales.
ACKNOWLEDGEMENT
REFERENCES
Hubbard E.F., Kilpartick F.A., Martens L.A., Wilson J.F.Jr. (1982) Measurement of time of
travel and dispersion in streams by dye tracing. Techniques of Water-Resources Investigations
of the US Geological Survey, Book 3: Applications of hydraulics, pp 44.
HGI Geological Survey of Croatia (2011) Hydrogeological Map of Croatia, 1:300 000,
Zagreb 2011.
Kresic N., Z Stevanovic (2010) Groundwater Hydrology of Springs Engineering, Theory,
Management, and Sustainability. Elsevier, ISBN: 978-1-85617-502-9
McCuire K.J., McDonnell J.J. (2006) A review and evaluation of catchment transit time mod-
eling. Journal of Hydrology 330, 543563.
Pavuza R. & Traindl H. (1983) ber Dolomitkarst in sterreich. Die Hhle 34(1), 1525, 7
Abb., Wien.
Savic D., Dozed S. (1985) Osnovna Geoloka Karta SFRJ Delnice 1:100 000, Geoloski Zavod,
OOUR Geologiju i Paleontologiju, Zagreb, Geoloski Zavod Ljubljana.
Stadler H., Skritek P., Sommer R., Mach R. L., Zerobin W., A.H. Farnleitner (2008)
Microbiological monitoring and automated event sampling at Karst springs using LEO-
satellites, Water Science Technology 58(4), 899909.
Strobl T., Zunic F. (2006) Wasserbau, Aktuelle Grundlagen Neue Entwicklungen, Springer-
Verlag, pp 9295.
Vreca P., Bronic I.K., Horvaticic N., Bareic J. (2006) Isotopic characteristics of precipitation in
Slovenia and Croatia: Comparison of continental and marine stations. Journal of Hydrology
330, 457469.
This page intentionally left blank
Chapter 14
Scale-dependent evaluation
of an unconfined carbonate
system Practical application,
consequences and significance
dm Tth & Judit MdlSzo nyi
Department of Physical and Applied Geology, Etvs Lornd University,
Budapest, Hungary
ABSTRACT
Understanding of karst systems at basin scale has only been highlighted in a few theoreti-
cal studies. This paper intends to demonstrate the applicability of the gravity-driven regional
groundwater flow (GDRGF) concept in karst systems. For this purpose, the hydrogeological
conditions of a golf course, situated in the Balaton Highland, Hungary, were examined at
different scales. The goal was to define the appropriate scale and to reveal the effects of dif-
ferent structures; and to give prognoses for the possible impact of a planned drinking water
well on the golf course. Field-scale study has shown that the discharge at the golf course is
tectonically-controlled and it is fresh karst water. Cluster analysis of spring data resulted in
heterogeneity- and structure-related groups at aquifer scale. The real underground flow path of
spring groups could be interpreted only in GDRGF context at basin scale. The appropriate scale
to solve the conflicts could be derived.
14.1 INTRODUCTION
Karst-related studies nowadays have to react to the future climate prognoses and
deal with their consequences on water utilisation possibilities in these changing cir-
cumstances. The review of the most important approaches of karst modelling for
water resource predictions by Hartmann etal. (2014) has revealed the difficulties of
representing spatial heterogeneities in karst and the uncertainties due to restricted
information on observed discharges in the system. These can cause uncertainties also
in the evaluation of the effect of future climate changes on karst systems. The main
conclusions of the review were that we need better methods to apply karst models
at large scales and we could make progress in the large-scale application of karst
models (Hartmann etal., 2014 p.18). These conclusions clearly show the demand for
understanding the processes on larger spatial scale in carbonates.
Karst systems are mostly integral parts of sedimentary basins. Therefore, the
applicability of the gravity-driven regional groundwater flow (GDRGF) concept to
the unconfined and adjoining confined carbonate systems on basin scale (large scale)
can give a new insight into their processes as it was proposed by Mdl-Szo nyi and
Tth (2015) (Figure14.1). The application of this approach has proved that the effect
of heterogeneity and anisotropy on the flow pattern could be derived from hydraulic
responses of the system on this scale. In addition, this study demonstrated that the eval-
uation of natural springs, as discharge phenomena (Tth, 1971), reveals the nature,
200 Karst without Boundaries
unconfined confined
freshwater
wetland Ca2+, Mg2+, HCO3, SO42, Na+, Cl, 222Rn, 226Ra
river
12 km
some kms
Figure14.1 Conceptual Tth-type GDRGF pattern of thick carbonate systems and its consequences
on flow-related manifestations; the shallow karst aquifer (modified after Goldscheider and
Drew 2007) is embedded into the regional flow pattern as a local system (Figure 14.10a
of Mdl-Szonyi and Tth, 2015).
Scale-dependent evaluation of an unconfined carbonate system 201
The local study area, a golf course, is situated in the unconfined part of the
Transdanubian Range Unit, a few kilometres from Lake Balaton, Hungary
(Figures14.2ab). The conflicts of interest in the area occur because a karst drinking
water well is planned near (~1500 m) the golf course (ENVICOM2000, 2012). The
question is how will the natural discharge on the golf course be influenced by the
karst drinking water production well. The goal of this study is not to solve this local
problem. This chapter intends only to demonstrate the importance of the understand-
ing the appropriate scale in karst studies and to illustrate how the GDRGF concept
can help to determine it.
The Transdanubian Range (TR) of Hungary, extending for 250 km in a north east
south west direction, is bounded by strike-slip and normal faults and consists of geo-
logically complex hills and mountains (Haas 2001, 2012). The Transdanubian Range
Unit (Figure14.2a) is built up mainly of Triassic and Early Jurassic confined and
unconfined carbonates, but beside carbonates Paleozoic metamorphic and Permian
Cretaceous sedimentary rocks can be found also in the region (Haas, 2001). The
202 Karst without Boundaries
A)
unconfined carbonates
Cretaceous
Jurassic
Triassic
B)
Holocene siliciclastic sediments
Triassic carbonate aquifer
transverse fault
golf course
elevation (m asl)
Lake Balaton
104 m asl planned well
0 500 m wetland
Figure14.2 A) The Transdanubian Range Unit of Hungary, boundaries and distribution of Paleozoic
and Mesozoic formations and major structural elements (modified after Haas (2001) a fter
Flp etal. (1987)) in addition the delineation of unconfined and confined carbonates with
the different study scales. B) Topography, hydrostratigraphy and the main structural ele-
ments of the field-scale study area. Location of the golf course and position of the planned
water abstraction well are also indicated. Structures are adapted from Dudko (1991).
structural pattern is characterised by a large north east south west trending synform
accompanied by antiforms (Fodor, 2010; Haas, 2012).
The Transdanubian Range is the biggest karstified carbonate aquifer system in
Hungary (Csepregi, 2007). Natural discharge takes place via cold, lukewarm and
thermal springs as well as creeks and wetlands. Groundwater in the area is used
mainly for public water supply, healing and heating purposes, and also as mineral
water. The region is rich in brown coal and bauxite resources which can be found
below the water table, therefore, for economic production, mine dewatering (between
Scale-dependent evaluation of an unconfined carbonate system 203
the 1950s to 1990s) was required which has caused significant water level decrease
(Alfldi & Kapolyi, 2007).
The region of the golf course is located between 115135 m elevation. Originally,
it had gradually sloping relief but nowadays it is an artificially-formed hummocky
small plateau with several tiny artificial lakes all but one of which are lined (Figures
14.2b and 14.3). The climate is continental with some Mediterranean influence (aver-
age annual precipitation is ~500700 mm). The area of the golf course is characterised
by natural diffuse discharge. East and south of the golf course small natural creeks
can be found as discharge features. Historical records, memoirs and archaeological
artefacts indicate that this site has been inhabited permanently from the 3rd century
suggesting the area has had a constant and stable water supply for human beings
(ruins of Roman Ages can be found here around a spring [Figure14.3]) (http://www.
orvenyes.hu).
Geologically the area is built up by unconfined Triassic carbonates and they are
partially covered by Quaternary siliciclastic sediments (Figure14.2b). At this site
the Triassic carbonates can be distinguished according to their hydraulic behaviour.
Undifferentiated dolomites and limestones are the main aquifer for karstwater in
the region with hydraulic conductivity (K) of 10-510-4 m/s. However, the cherty and
tuffaceous carbonate formation has a lower K (~10-710-6 m/s), therefore, it functions
as aquitard (Tth etal., 2014). The covering siliciclastic sediments have limited exten-
sion and they might have similar hydraulic properties to the Triassic aquifer depend-
ing on grain size distribution and degree of compaction. However, the region is highly
influenced by tectonic events through geological evolution (Figure14.2b).
N golf course
spring
spring
spring
spring
wetland
RMT site
VLFEM profile
Lake
0 100 200 m
Balaton detected fault
The goal of the field-scale study at the golf course was to characterise the natural
discharge features. Water samples for chemical analysis were taken from the springs,
geophysical measurements were carried out, as well as field observations. These data
were complemented by archive data. The hydrostratigraphic and tectonic influence on
discharge and the chemical character of springs were determined.
There are four springs that discharge in the vicinity of and within the golf course
(Figure14.3). Beside discrete natural discharge points, diffuse discharge also occurs
forming a chain of waterlogged areas usually in local depressions of pot-holed terrain
and around artificial lined lakes. These groundwater conditions together with the
archive data prove the continuous existence and stability of water discharge in the
area, a setting that is favourable for maintaining the fresh grass vegetation of the golf
course.
Water samples of springs (indicated by circle, triangle, pentagon and asterisk) and
lake without lining (indicated by grey dashed area) (Figure14.3) were collected to
assess the chemical character of the discharges. The low TDS values (760835 mg/l)
of all the water samples indicate the existence of fresh karst water characterised by
calciummagnesium and bicarbonate facies (Table14.1, Figure14.4). The highest
TDS occurs in the spring (displayed by an asterisk) closest to the Lake Balaton.
These parameters represent relatively a shallow flow path and the magnesium
concentration reflects interaction with dolomite. According to the archive dataset
(Hungarian Spring Cadastre) the spring volume discharge rates in this area vary
across a wide range with 0.160.33 l/s median values of discharge rate.
Electromagnetic geophysical measurements were carried out to determine and
specify the hydrostratigraphic structure and see whether it has a modifying effect on
the discharge pattern. Electromagnetic techniques are efficient tools for investigating
shallow (~100 m) geological formations, and are commonly applied in hydrogeolog-
ical research (e.g. Turberg etal., 1994; Nobes, 1996; Meju etal., 1999; Vereecken
etal., 2005) because geological sequences (aquitards and aquifers) and structures (e.g.
faults, fractures) can be investigated quickly, extensively and effectively (Tezkan etal.,
2000; Gurk etal., 2001; Bosch & Mller, 2001). Radio-magnetotellurics (RMT) was
applied to provide information about hydrostratigraphy and Very Low Frequency
Electromagnetics (VLFEM) to detect tectonic elements and fractures, respectively
(Figure14.3).
80
80
Ca
60 (Cl)
lciu
id e
m
(C
lor
60
a)
40 4 + Ch
+ Ma
)
gne
SO
40
siu
e(
m
fat
(M
Sul
)g
20
20
Mg SO4
20
O 20
)
20 3
HC
20
80
80
40
40 m (N
So
e(
diu
nat
rbo
g)
(M
Sul
ica
60
40
40
60
a)
fat
60 3 ) + B
m
O 0
60 tassi
+P
6
siu
e(
o
gne
SO 4
(C
Ma
)
um
40
40
60
e
nat
rbo 0
80
8
(K)
Ca
80
20
20
80
Ca 80 60 40 20 Na+K HCO3+CO3 20 40 60 80 Cl
Calcium (Ca) Chloride (Cl)
% meq/l
CATIONS ANIONS
Figure14.4 Piper diagram of water samples of the golf course (for location of data points see Figure 14.3).
SW NE
130
33 30
139
366 13 17
120 64
33
e
347
elevation (m asl)
chang
386 33
110
79
gical
39
litholo
100 28
26
65 59
sharp
65
90 95
822 65
80
0 150 200
golf course
strike-slip fault
N
transverse fault
Lake Balaton
detected fault
supposed tectonic line
artificial lake without lining
spring
spring
spring
spring
ruins of Roman Ages
0 100 200 m wetland
planned well
golf course cannot be shown, however, these questions can be answered by expanding
the study area and lowering the scale.
In the next phase, the study area was expanded to a larger, aquifer-scale area (10skm
10skm) of the Balaton Highland (Figure14.2a). The delineation of the area was
based on topographic considerations and the structural pattern, because the north
west thrust fault is supposed to separate the Balaton Highland from the more elevated
part of the Transdanubian Range (Csepregi, 2007). To understand the flow systems to
the discharges at the golf course, the natural springs of the area were used to provide
information about their parent flow systems. For this reason, archive data of springs
(based on Hungarian Spring Cadastre) were examined and analysed by a descriptive
statistical method to better understand the hydrogeological systems.
Cluster analysis (e.g. Kovcs etal., 2012) was used to determine the possible groups
of springs. The archive spring database of the Transdanubian Range was used for the
basic data regarding springs, such as elevation, volume discharge, temperature and the
most important water chemistry. The chemical parameters reflect different orders of
flow systems that can be represented in the spring groups (Mdl-Szo nyi & Tth, 2015).
The natural spring data were influenced by artificial mine dewatering (Alfldi &
Kapolyi, 2007) therefore, only the data before 1960 were used to examine the natural
flow conditions. The first parameter was the elevation of the spring orifice, the second
one was the water temperature because there are cold and also lukewarm springs. The
third parameter was the chloride content because it is a conservative natural tracer
in water. Finally, the volume discharge of the springs was the fourth parameter, since
springs fed by local flow systems display higher variability in discharge while higher
order of flow systems are more stable (Bodor etal., 2014; Mdl-Szo nyi & Tth, 2015).
The results of the cluster analysis show that the groups are more or less separated
by elevation of spring orifice and water temperature (Figure14.7). However, there are
no significant changes in chloride concentration, discharge rates vary widely, there-
fore these two parameters cannot be applied for distinguishing spring groups.
The group of springs delineated by a triangle in Figure14.7 are characterised by
the lowest discharge elevation (135 m asl) and the highest temperature (12.5C). The
springs, indicated by a circle, discharge at higher elevation (~200 m asl) but are char-
acterised by a consistent temperature of 11C. The areal distribution of the other two
groups does not show any definite pattern which can be related to groundwater flow
systems. The springs on the field scale were very similar, with almost the same chem-
ical composition, here they are part of different groups. Nevertheless, the position of
springs are in good correlation with tectonic elements of the Balaton Highland area,
i.e. on the aquifer scale the springs are separated into different groups supposedly due
to the effect of geological heterogeneities and structures.
The origin of springs in this larger study area, the aquifer scale, have not been
determined. Clustering the springs of the Balaton Highland region resulted in a geo-
logical heterogeneity-related grouping without any information about the parent
groundwater flow systems. For this reason, the study area was further expanded to
the BakonyBalaton Highland region.
A)
ton
ala
eB
Lak
B) 400 20
Temperature (C)
Elevation (m asl)
300
10
200
100 0
Discharge (l/min)
80 800
Cl (mg/l)
40 400
0 0
Figure14.7 A) Areal distribution of spring groups on aquifer scale for the Balaton Highland (sur-
rounded by dotted line). Dashed circle indicates the vicinity of field-scale study area.
Tectonic elements of the Balaton Highland are adapted from Dudko (1991), transverse
faults are indicated by grey colour. B) The box-and-whisker plots of the parameters.
Scale-dependent evaluation of an unconfined carbonate system 209
The same statistical data analysis was done for the bigger study area (some 50 km
some 50 km) whih included the Bakony Mountains and Balaton Highland. In this case
the parameters were more indicative of the hydrogeological settings (Figure14.8).
According to the elevation of the spring and the water temperature the springs form
four distinctive groups: springs with relatively high water temperature can be found at
low elevations around the BakonyBalaton Highland area, while the coldest springs
are situated at the elevated parts of the Bakony Mountains. The discharge rate of the
springs still varies in a large range suggesting that there are springs with stable and
high amounts of discharge and others with fluctuating discharge rates which may
even dry up altogether. Variation of the chloride ion concentration is not significant
because this region is unconfined and contribution from siliciclastic confining strata
is small (Mdl-Szo nyi & Tth, 2015).
These groups can be interpreted in the GDRGF context. The group indicated
by triangles (Figure14.8) represents the springs of the highest Bakony, above the
Cretaceous marl in perched carbonate aquifers. Groups represented by circles and
squares discharge at lower elevation (~150300 m asl) with a slightly elevated tem-
perature (~1013C). They relate to two different local flow systems. The group pre-
sented as asterisks has the highest temperature and it is characterised by the lowest
elevation with a significant discharge rate (~8.33l6.66 l/s). These springs can be
identified as fed by an intermediate flow system. In this basin-scale study the areal
pattern of springs are more consistent with the topography-driven groundwater flow
conditions. By lowering the scale, i.e. expanding study area, the areal pattern of the
springs does not alter as demonstrated by Mdl-Szo nyi & Tth (2015) for the whole
Transnubian Range.
The relationship between spring groups of the field-scale study area and the simu-
lated flow pattern were compared based with the scenario simulations of Mdl-Szo nyi
& Tth (2015, Figure 14.2a). The springs of the field-scale study indicate the same,
supposedly local flow-related group (squares on Figure14.7). Consequently, this scale
is necessary to explain the source and flowpath of groundwater towards the golf
course area, since GDRGF is the most important agent within this framework.
In the simulated section the northern boundary of the source area of the golf
course a low permeability thrust fault is present (Figure14.7). However, the simulated
flow pattern displays that there is throughflow across this fault in spite of its low per-
meability, because the main hydraulic gradient is perpendicular to the fault. There is
a deeper throughflow component as well, which can contribute with a deeper water
component to the springs (the asterix group in Figures 14.3 and 14.4.).
14.8 DISCUSSION
The direct recharge area of the golf course can be found down-dip of the thrust fault
between the Balaton Highland and the Bakony Mountains (~12 km in Figure14.9).
This direct recharge produces a large amount of water flowing towards the golf course.
However, there is some throughflow across the thrust fault, from higher topographic
regions (this additional recharge is at ~8 km) and a deep water flow contribution as
A)
ton
ala
eB
Lak
B) 500
20
Elevation (m asl)
Temperature (C)
400
300
10
200
100 0
8000
Discharge (l/min)
80
Cl (mg/l)
4000
40
0
0
Figure14.8 A) Areal distribution of spring groups on basin scale for the whole Bakony Mountains and
Balaton Highland. Dashed circle indicates the vicinity of field-scale study area and dotted
line surrounds the aquifer-scale study area. B) The box-and-whisker plots of the parameters.
Scale-dependent evaluation of an unconfined carbonate system 211
500
Elevation
(m asl)
0
500
1000
0 5 10 15 20
Distance (km)
Figure14.9
Flow distribution, heads (m), streamlines and Darcys velocity vectors (magnitude-
controlled) based on flow simulations along a section (see location on Figure 14.2a) of
Mdl-Szonyi & Tth (2015; Figure 14.7c) and with the indication of the supposed recharge
and discharge area of golf course and the projected location of the planned well.
well (Figure14.9). The velocity vectors are displayed in magnitude, therefore, they
do not represent the drop in discharge across the fault. The clustering of springs on
the basin scale represents local flow systems (Figure14.8) with CaMg and HCO3
water types. Only one spring displays a higher TDS and lower magnesium and higher
sulphate and chloride concentrations, indicating some additional deeper water contri-
bution (Table14.1, Figure14.4). The south east boundary of the discharge region of
the golf course is the strike-slip fault (Figures 14.2b and 14.6). The discharge of the
spring with elevated TDS (indicated by asterix) can be connected to this. There are
two transverse faults which influence the delineation of the study area. The north east
and south west boundary of the study area is a stream and the western transverse fault
respectively (Figure14.2b). Due to limited throughflow from the Bakony Mountains
across the thrust faults and the efficient hydraulic continuity in the carbonates, the
effect of the planned production well on the discharge at the golf course cannot be
overlooked. The strike-slip fault in the south east and the thrust fault in the north
west, as hydraulic boundaries, may contribute to this effect based on the supposed
superposition of drawdown.
3D transient numerical simulation of the potential hydraulic effects of the new
well require the field-scale study to be extended as far as the western transverse fault
(Figure14.2b) to understand the hydrostratigraphy and tectonics of the area between
the well and the golf course.
14.9 CONCLUSION
The objective of this study was to demonstrate the usefulness of the GDRGF con-
cept in solving karst-related problems in an unconfined carbonate basin. Theoretical
research suggested that karst systems, as integral parts of sedimentary basins are
characterised by gravity-driven flow of adjoining confined and unconfined carbonate
systems. This study demonstrated the importance of springs and connected natural
212 Karst without Boundaries
discharge features to reveal the nature, hierarchy, chemistry and temperature distri-
bution of topography-driven regional groundwater flow. Hydraulic connection also
occurs between large, basin-scale carbonate and shallow aquifer-scale karst features.
On regional scale, hydraulic continuity or compartmentalisation of karst regions by
structures or faults are more significant than the effects of local heterogeneities. In
addition to these theoretical findings the applicability of GDRGF approach to solve
practical questions in carbonate regions need to be assessed. One of the most import-
ant practical questions is the delineation of the appropriate study areas in such regions
and to learn more about the effects of faults and hydraulic connectivity between dif-
ferent parts of the carbonate basin.
ACKNOWLEDGEMENTS
The authors appreciate the help of collaborating colleagues (Imre Mller, Szilvia
Simon, Ferenc Zsemle, Tmea Havril) and students during field-scale study. Technical
support of Jzsef Kovcs, Gergely Hornyk, Petra Bodor and Soma Szathmri is also
acknowledged. The access to the field-scale problem was provided by Tibor Sarls and
Jzsef Vers. Valuable comments of anonymous reviewers are appreciated. Language
editing by Nicholas Robins is highly acknowledged.
REFERENCES
ABSTRACT
Hydrogeological models are an approximation of real systems. The structure of these models,
includes several equations which are known as the system transfer functions that define the
relationship between inputs and outputs. The transfer functions and the impulse response func-
tions (i.e., Kernel function) of the karst springs, Pireghar and Dimeh, are identified by means
of the ARX model. Daily rainfall data of the Chelgard and Farsan rain gauge stations and
daily spring discharge of the Dimeh and Pireghar Springs during the period of 19992011 are
selected as input and output, respectively in the modelling processes. The results show different
contribution of the flow components, slow flow and quick flow, in the springs. The slow flow
component in the Dimeh and Pireghar springs is 90% and 64% of the total spring discharge,
respectively. Moreover, in order to evaluate spring discharge variations in response to a hypo-
thetical drought conditions (assuming a decreasing trend in precipitation during the next twelve
years), the spring discharge is predicted based on the calibrated model. The results reveal that,
the Pireghar Spring will be more affected by drought events compared to the Dimeh Spring.
15.1 INTRODUCTION
Karst aquifers contain significant water reserves in many parts of the world, but
development and management of these resources is difficult because of the complex-
ity of the aquifers. Karst aquifers are mainly characterised by an irregular network of
pores, fractures, fissures and solution conduits in various sizes and shapes. Such struc-
tures with inherent heterogeneous geometry cause complex hydraulic conditions and
spatial and/or temporal variability of hydraulic parameters in these aquifers. After a
rainfall event, water primarily recharges through large conduit, fissures and fractures
and cause a large amount of water moves rapidly to the outlet of karst spring. But in
smaller fractures and fissures, slow and predominantly laminar recharge occurs due to
gradual drainage from them. These two processes form a quick flow and a slow flow
component of a karst spring hydrograph (Denic-Jukic & Jukic, 2003).
Assuming a karst aquifer as linear, time-invariant and casual system, discharge of
the karst spring at a time t may be represented by the superposition of three compo-
nents as shown in Figure 15.1 (Denic-Jukic &Jukic2003):
Where yD(t), yS(t) and yQ(t) are named as a component of antecedent flow, slow
flow and quick flow, respectively (Figure15.1).
yD(t) is a component of the spring discharge resulting from antecedent rainfalls.
It can be defined by using the superposition law and the property of independence
of discharge events. It practically means that the component of discharge resulting
from antecedent rainfall is equal to the spring discharge resulting of a period without
rainfall. In this regard, the antecedent recession curve has the same form as the master
recession curve. Consequently, the function yD(t) has the exponential form with the
recession coefficient :
yD(t) = y0e-t/(15.2)
where y0 is the karst spring discharge at the beginning of the period analysed.
yS(t) belongs to the water storage in the pore space and micro fissures which
defined as the base flow. Generally groundwater flow in this part of the karst media
known as the diffuse flow regime.
yQ(t) relates to quick recharge through sinkholes, conduit and large fissures and
movement of groundwater in the karst system under the dominant turbulent flow
regime.
The discharge component of yD(t) is the result of an initial storage in the karst
system, while yS(t) and yQ(t) are the result of current rainfall (e.g., x(t-), [0,t]).
The relationship between the rainfall and the resulting karst spring discharge can be
represented by the linear convolution integral between the rainfall and a transfer func-
tion, so Eq. 15.3 may be written as (Denic-Jukic & Jukic, 2003):
t t
y(t) = y D (t) + hS ( )x(t - )d + hQ ( )x(t - )d (15.3)
0 0
where hS() and hQ() are the slow flow and quick flow transfer function, respectively.
One of well-known methods for characterising karst aquifers is modelling of
the spring hydrograph based on the use of kernel functions (i.e., impulse response
function). When the kernel function of a karst system is known, the output (i.e.,
spring discharge) can be determined for any specified input (i.e. rainfall events). If
the internal structure of the system is known, the response function can sometimes
Antecedent
Antecedent rainfalls flow function
Figure15.1 The superposition of three components to produce the spring discharge water.
Characterization of karst system using modelling of rainfall-discharge relationship 217
The study area is located in the Zagros Mountains Range, which comprises a series
of parallel north west-south east trending anticlines and synclines formed by com-
pressional tectonics during the Miocene age. The stratigraphic and structural setting
of the Zagros Mountain range is described in detail by Stocklin & Setudehnia (1977)
and Alavi (2004). The Alpine Orogeny led to many thrust faults in the area. The
major geological formations in the study area (Figure15.2), in increasing order of
age, include Pabdeh-Gurpi Formation (Palaeocene/Oligocene), Asmari-Jahrom and
Shahbazan Formation (Oligocene to early Miocene), Razak Formation (Miocene),
Gachsaran Formation (Miocene-Eocene), Mishan Formation (Middle to Late
Miocene), Aghajari Formation (late Miocene to Pliocene), Bakhtiari Formation (late
Pliocene-Pleistocene) and the recent alluvium. The Asmari Formation has the poten-
tial to form a karstic aquifer system. The Asmari-Jahrom and Shahbazan Formations
includes limestone layers, thin layers of clay and gray limestone along with the fossils
orbitolina and shale, marl and sandstone of Cretaceous age. The Upper Cretaceous
and Tertiary geological formations comprise limestone layers which are suitable for
the karst development.
The Asmari-Jahrom and Shahbazan formations outcrops mainly around Dime
and Pireghar springs and the Saldoran and Zarab Mountains in this area are probably
part of the catchment area of the springs (Figure15.2). Large dissolution cavities have
been observed in the massive limestone rocks of the Asmari-Jahrom and Shahbazan
formations in the area, which could be considered as indications for development of
karst. Different karstic phenomena such as springs, caves (e.g. Pireghar and Sarab
Spring
N
Theran
Pireghar spring
Shahrekord
Study area
Pe
rsi
an 0 5 15 km
Gu
lf
5015 5030
caves), karren, dissolution cavities, poljes, dolines, and dry valleys are observed
locally. Several joint sets were extensively developed due to the Zagros Thrust Fault
Zone, which provides favourable condition for surface recharge in the carbonate
rocks consequent to the development of the karstic features.
The Dimeh and Pireghar springs are located in Zagros Region, Iran. A catchment
area of 140 and 90 km2 were determined for the Dimeh and Pireghar, respectively (Mali,
2014). Mean annual rainfall in the catchment area of the Dome and Pireghar springs
is 542.5 and 1329.1 mm based on data from Farsan and Chelgerd rain gauge stations
during the period of 19992011, respectively. Spring discharge in the Dimeh and Pireghar
springs range from 1.5 to 4.3 and from 0.6 to 5.3 m3/s, respectively. The ratio of max-
imum to minimum spring discharge were computed 2.9 and 9.58 for the Dimeh and
Pireghar springs. Mali & Mohammadi (2013) have shown that the recession coefficient
value () of the Dimeh and Pireghar Springs ranges from 10-3 to 10-4 and 10-2 to 10-3.
15.3 METHOD
The ARX model is one of the most popular models for modelling time-invariant linear
systems. This model is one of the hybrid models (i.e., Gray box model) which charac-
terise by specifying structure. Parameters of the ARX model are determined via cali-
bration and verification process. According to the ARX model, any spring discharge
record, y(k), at time k is dependent on values of simultaneous rainfall, x(k), and prior
rainfall and spring discharge records as follows:
B(z -1 ) 1
Y(z) = -1
X(z) Z - b + E(z) (15.7)
A(z ) A(z -1 )
where
The polynomial B(z-1) is the impact on the spring discharge from antecedent
rainfall events and spring discharges. When coefficient n is large, it indicates that
220 Karst without Boundaries
yt = ht xt = hx
k =-
k t -k
(15.10)
5 6
R = 0.93 5
Calculated
4 R = 0.96
Calculated
RMSE = 0.007 4
3 RMSE = 0.04
3
2 2
1 1
(a) (b)
0 0
0 1 2 3 4 5 0 1 2 3 4 5 6
Observed Observed
Figure15.3 Observed and calculated discharge for (a) Dimeh Spring and (b) Pireghar Spring.
Characterization of karst system using modelling of rainfall-discharge relationship 221
between the rainfall and the spring discharge in the Pireghar Spring is probably non-
linear and the ARX model could not exactly simulate the transfer function of the karst
system of the Pireghar Spring.
Figure15.4 shows temporal variations of the observed discharges of the Dimeh
Spring and the three components of spring discharge, including antecedent, slow
and quick flow, calculated by the selected ARX model in the verification step. The
antecedent recession curve decreases with an exponential trend (yD(t)=y0e-0.001t). It
means that with a lack of recharge from rainfall, the spring will dry out during the
next 30003500 days. The slow flow contributes to 90% of the total spring discharge
and the contribution from the quick flow in the total spring discharge is low.
The recession trend of the antecedent flow in the Pireghar Spring shows the
spring will dry out during the next 250300 days, with no recharge from rainfall
(Figure15.5). The contribution of the slow flow is about 63% of total spring dis-
charge in the Pireghar Spring (Figure15.5). Contribution of the quick flow in the
total spring discharge in the Pireghar Spring is greater than that in the Dimeh Spring.
6
observed discharge quick flow component
5 slow flow component antecedent component
Discharge (m3/s)
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Time (year)
Figure15.4 Temporal variations of observed discharges of Dimeh Spring and flow components as
antecedent, slow and quick flow.
12
observed discharge quick flow component
slow flow component antecedent component
10
Discharge (m3/s)
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Time (year)
Figure15.5 Temporal variations of observed discharges of Pireghar Spring and flow components as
antecedent, slow and quick flow.
222 Karst without Boundaries
The response of the springs to rainfall events is quite different. The impulse
response function (i.e. Kernel function) in the Pireghar Spring is abrupt with a sharp
peak and shows less retention time in comparison to the Dimeh Spring (Figure15.6).
According to the shape of kernel functions, it seems that groundwater flow in Pireghar
Spring and the Dimeh Spring is mainly controlled by the conduit and diffuse flow
regimes, respectively (Figure15.6). The rising and falling limb of the Kernel function
in the Pireghar and Dime springs have different slopes due to different hydrogeolog-
ical characteristics such as degree of karst development and/or size and shape of the
catchment area of the springs.
Variation of the springs discharge in response to a hypothetical drought condi-
tions (e.g., assuming a decreasing trend for rainfall during the next twelve years) is
evaluated using predicted values by the calibrated models for the Pireghar and Dime
springs (Figures 15.7 and 15.8). Discharge of the Dimeh and Pireghar springs has
decreased to one third and one fifth of the mean annual discharge after about 8 years,
respectively. Comparison of the spring discharge under assumed drought conditions
with the current spring discharge reveals that discharge of the Pireghar and Dimeh
springs decreases bb 59% and 36%, respectively. It seems that drought crisis is more
0.012
Pireghar spring Kernel
0.01
Dimeh spring Kernel
0.008
h(t)
0.006
0.004
0.002
0
0 200 400 600 800 1000 1200
Time (day)
Figure15.6 Impulse response function (Kernel function) in Pireghar and Dimeh springs.
6 0
50
5 Observed discharge 100
Discharge (m3/s)
Simulated discharge
4 150
Predicted discharge
200
Rainfall (mm)
Rainfall
3 Assumed rainfall
250
300
2
350
1 400
450
0 500
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Verification period Prediction period
Time (year)
Figure15.7 Calculated discharge of Dimeh Spring in response to real and assumed rainfall events.
Characterization of karst system using modelling of rainfall-discharge relationship 223
0
14
12 50
Observed discharge
Discharge (m3/s)
Rainfall (mm)
10 Simulated discharge 100
8 Predicted discharge
150
Rainfall
6 Assumed rainfall
200
4
2 250
0 300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Verification period Prediction period
Time (year)
Figure15.8 Calculated discharges of Pireghar Spring in response to real and the assumed rainfall
events.
obvious in the Pireghar Spring with the dominant conduit flow regime than the Dime
Spring with its diffuse flow regime. A karst spring with a conduit flow system is more
vulnerable to drought.
15.5 CONCLUSIONS
The contribution of the slow flow in the discharge of the Pireghar and Dimeh springs
is about 64% and 90% of the total spring discharge, respectively. The impulse
response function in Pireghar Spring is abrupt with a sharp peak and includes a smaller
retention time in comparison to the Dimeh Spring. It seems that the karst is probably
developed to a lesser degree in the catchment area of the Dimeh Spring compared to
that of the Pireghar Spring. Prediction of spring discharge reveal that the Pireghar
Spring will be more vulnerable to drought crises than Dimeh Spring. In conclusion,
the diffuse flow system in the catchment area of the Dimeh spring helps with storage
and gradual releasing of stored groundwater during drought periods.
Comparison of the developed kernel functions for the springs confirms the con-
duit and diffuse flow regime in the catchment area of the Pireghar and Dime Springs,
respectively. Lag time between rainfall events as input to the model and spring dis-
charge as model output in the Pireghar Spring is lower than the Dimeh Spring. Even
though the ARX model that was applied is a linear model, according to the results, it
seems that the relationship between rainfall and spring discharge in both springs are
not completely linear. In order to obtain a more exact simulation and prediction, a
nonlinear model such as NARX is suggested.
ACKNOWLEDGMENTS
This research was supported financially by the Research Council of Shiraz University.
The authors would like to thank the Chaharmahal and Bakhtiari Regional Water
Authority for providing data.
224 Karst without Boundaries
REFERENCES
Alavi M. (2004) Regional stratigraphy of the Zagros fold-thrust belt of Iran and its pro-
Foreland evolution. American Journal of Science 304, 120.
Barrett M.E., Charbeneau R.J. (1997) A parsimonious model for simulating flow in karst aqui-
fer. Journal of Hydrology 196, 4765.
Denic-Jukic V., Jukic D. (2003) Composite transfer functions for karst aquifers. Journal of
Hydrology 274, 8094.
Eisenlohr L., Bouzelboudjen M., Kiraly L., Rossier Y. (1997) Numerical versus statistical mod-
elling of natural response of a karst hydrogeological system. Journal of Hydrology 202,
244262.
Fazeli M.A., 2007. Construction of grout curtain in karstic environment case study: Salman
Farsi Dam, Iran. Environmental Geology 51, 791796.
Halihan T., Wicks C.M. (1998) Modelling of storm responses in conduit flow aquifers with
reservoirs. Journal of Hydrology 208, 8291.
Labat D., Ababou R., Mangin A. (1999) Linear and nonlinear input/output models for karstic
spring flow and flood prediction at different time scales. Stochastic Environmental Research
and Risk Assessment 13(5), 337364.
Mali S. (2014) Prediction of karst aquifer behaviours by rainfall-spring discharge analysis: Case
of Pireghar and Dime Springs, Shahr Kord M.S Thesis, Shiraz University, Shiraz.
Mali S., Mohammadi Z. (2013) Investigation of karstification by hydrograph and time-series
analysis. The 17th Symposium of Geological Society of Iran and the 1st International
Congress on Zagros Orogen Shahid Beheshti University, Tehran, Iran, Oct. 2931.
Neuman S.P., de Marsily G. (1976) Identification of linear systems response by parametric
programming, Water Resources Research 12(2), 253262.
Oppenheim A.V., Willsky A.S. (1997) Signal & System. 2nd ed, Prentice Hall.
Singh V.P. (1988) Rainfall-runoff Modelling, Hydrologic Systems, vol. 1. Prentice-Hall,
Englewood Cliffs, NJ.
Stocklin J., Setudehnia A. (1977) Stratigraphic Lexicon of Iran. Geological Survey of Iran
Tehran, Iran, 376 pp.
Teutch G., Sauter M. (1997) Distributed parameter modelling approaches in karst h ydrological
investigations. Proceedings of the Sixth Conference on Limestone Hydrology and Fissured
Media, La Chaux de Fonds, Switzerland.
Part 3
ABSTRACT
The recharge processes have been evaluated for several main karst massifs of southern Italy,
including the Terminio, Cervialto and Matese massifs, characterised by wide endorheic areas.
The annual mean recharge has been estimated with GIS-basedtools, from regression of annual
mean values of different ground-elevated rain gauges and thermometers. The recharge has been
separated from endorheic areas and the other areas of spring catchment, and the ratio between
the output spring and input rainfall has been estimated (the recharge coefficient). Endorheic
areas are exploited for hydro-electrical purpose, especially in the Matese massif, and several
hydraulic works have modified the natural recharge processes and the regime of karst springs.
16.1 INTRODUCTION
Along the Apennine chain, many karst massifs outcrop, which are generally hydrauli-
cally isolated by thick flysch sequences or other non-karstic rocks along their bound-
aries. These massifs constitute the main water resources of central and southern Italy,
and feed several powerful karst springs.
Locally, due to the Mediterranean climate, the recharge generally occurs during
autumn and winter; as a consequence, karst aquifers discharge water throughout
the hydrological year, primarily in response to this climatic regime, but also to the
geological-geomorphological setting and karst conditions (Fiorillo, 2009).
The recharge processes of Terminio, Cervialto and Matese massifs, (Southern
Italy), are characterised by wide endorheic areas. Endorheic areas are closed depres-
sions, generally characterised by a seasonal or permanent sinking lake located at the
bottom; in these areas the runoff is completely adsorbed (White, 2002). The endor-
heic areas constitute important recharge zones of aquifers, generally hydraulically
connected to one or more springs; practically, all the rainfall falling on these areas
(net of evapotranspiration) infiltrates into the aquifer. Outside endorheic areas, part
of the rainfall can leave the spring catchment by runoff processes, especially during
intense storms, and are here referred to as open areas. The open areas constitute the
difference in area between the spring catchments and the endorheic areas.
Even if karst rocks allow a high rate of infiltration, runoff occurs especially along
steep slopes and during high rainfall intensity events. Volumes are difficult to esti-
mate, because they vary mainly as a function of several parameters such as the soil
moisture conditions and the intensity and distribution of the rainfall; besides, snow
accumulation and consequent snowmelt can have an important role in the runoff
228 Karst without Boundaries
These massifs are primarily characterised by steep slopes and high elevations up to
2050 m a.s.l. for Matese massif and 1809 m a.s.l. for Picentini mountains (Figure16.1).
Flat zones are limited to the bottom of endorheic areas which induce higher and more
concentrated water infiltration.
These karst massifs are made-up of a series of limestone and limestone-dolo-
mite (Late Triassic-Miocene), which are between 2500 and 3000 m thick. Along the
northern and eastern sectors, these massifs are tectonically overlapped on the terrige-
nous and impermeable deposits, constituting argillaceous complexes (Paleocene) and
flysch sequences (Miocene). Along the southern and western sectors, these massifs
are limited by normal faults and are covered by recent quaternary deposits. More
detailed geological desriptions of the outcropping areas can be found in Parotto &
Praturlon (2004) and related literature, and recent Geological Map of Italy, 1:50 000
scale (ISPRA, 2015).
Pyroclastic deposits of Somma-Vesuvius activity cover the Picentini mountains,
with thickness up to several meters along the gentle slopes of the Terminio zone, and
of few tens of meters along the steep slopes and the entire area of Mount Cervialto.
These deposits, almost absent on the Matese massif, play an important role in the
infiltration of water into the karst substratum.
A)
B)
Figure16.1 A) Study region in southern Italy B) Map of the Western Campania region. Rectangular
shaded area is detailed in figures16.2 and 16.3. Triangles are thermometer/rain gauges used.
The Serino group is located in the valley of the Sabato River, along the north west-
ern boundary of the Picentini massif and it is formed by the Acquaro-Pelosi springs
(377380 m a.s.l.) and the Urciuoli spring (330 m a.s.l.). These springs are fed by
Terminio massif (Civita, 1969, Fiorillo etal., 2007) with an overall mean annual
discharge of 2.25 m3/s. Local monitored piezometers have an upwelling flux in the
alluvial deposits and flysch sequences, which comes from the below karst substratum
of the Sabato valley, and supplies the springs. Roman aqueducts (first century AD.)
were supplied by these springs and the Urciuoli spring was re-tapped between 1885
and 1888 by the Serino aqueduct, which is a gravity channel followed by a system of
pressured conduits that is used to supply water to the Naples area. Additionally, the
Aquaro and Pelosi springs were also re-tapped in 1934 by the Serino aqueduct.
The Cassano group is located in the Calore river basin along the northern bound-
ary of the Picentini Mountains, and is formed by the Bagno della Regina, Peschiera,
Pollentina and Prete springs (473476 m a.s.l.). These springs are primarily fed by the
230 Karst without Boundaries
Terminio massif (Civita, 1969), with an overall mean annual discharge of 2.65m3/s.
In 1965 these springs were tapped to supply the Puglia region with water, and a grav-
ity tunnel was joined to the Pugliese aqueduct.
The Caposele group is formed by the Sanit spring (417 ma.s.l.), which is located
at the head of the Sele river basin along the north eastern boundary of the Picentini
Mountains (Figure16.2). This spring, which is primarily fed by the Cervialto moun-
tain (Celico & Civita, 1976), has a mean annual discharge of 3.96 m3/s. The spring
was tapped in 1920 by the Pugliese aqueduct, which passes through the Sele-Ofanto
divide via a tunnel and supplies the Puglia region with water.
The powerful springs of Matese massif are located along the southern and north-
ern side of the massif (Figure16.3). The Torano spring (201 m a.s.l.), located along
the eastern cliff of a karst canyon, is fed by the karst system of the central sector
of Matese massif (Civita, 1969). It has a mean annual discharge of 2.04 m3/s. The
Maretto spring (180 m a.s.l.) is located at the foot of limestone-dolomite slope, where
there is a contact with less permeable debris, which covers the limestone rocks, and
has a mean annual discharge of 0.99 m3/s. The Torano and Maretto springs were
tapped during the 1960s by the Campano aqueduct, that supplies the Naples area.
The Grassano-Telese springs, are located along the southern side of the Mount
Pugliano relief, near Telese village; this relief is separated from the Matese massif
(Figure16.3), but the springs have to be connected hydraulically with the wide area
of the Matese karst system (Fiorillo & Pagnozzi, 2015). Some of them are highly min-
eralised, and the total annual mean spring discharge is 4.3 m3/s.
1) 3) 5) 7) 9)
2) 4) 6) 8) 10)
0 1 2 3 4 5
Kilometers
Figure16.2 Hydrogeological sketch of north-eastern sector of Picentini Mountains (Fiorillo etal. 2015).
1) Slope breccias and debris, pyroclastic, alluvial and lacustrine deposits (Quaternary);
2) argillaceous complex and flysch sequences (PaleogeneMiocene); 3) calcareous-
dolomite series (JurassicMiocene); 4) main karst spring; 5) village; 6) mountain peak;
7) elevation (m a.s.l.); 8) Cervialto and Terminio groundwater catchment; 9) endorheic
area; 10) Caposele river gauge.
Recharge processes of karst massifs in southern Italy 231
Along the Northern side of the Matese massif, near the village of Boiano
(Figure16.3) there are three main groups of springs (Civita, 1969): the Maiella group,
west of Boiano, the Pietrecadute group located in the village, and Rio-freddo group,
east of Boiano. These springs are fed by the karst system of the north-central sector
of Matese massif and are located along debris deposits that cover the tectonic con-
tact between limestone and flysch sequences. Altogether these springs have an annual
discharge of about 2.80 m3/s. Most were tapped by the Biferno Aqueduct during
the 1960s and supply the Molise region. A branch of this aqueduct is joined with
Campano Aqueduct by a tunnel in the eastern side of Matese massif.
Other springs are located inside endorheic areas or in the high elevated zones, and
their discharge from the massif is lower than from the basal springs.
1) 6)
2)
7)
3) 8)
4) 9)
0 1 2 4 6 8 10
5) 10) Kilometers
Figure16.3 Hydrogeological sketch of Matese massif (modified from Fiorillo and Pagnozzi, 2015).
1) Slope breccias and debris, pyroclastic, alluvial and lacustrine deposits (Quaternary);
2) argillaceous complex and flysch sequences (PaleogeneMiocene); 3) calcareous-
dolomite series (JurassicMiocene); 4) main karst springs, 1 Grassano-Telese, 2 Boiano,
3 Torano, 4 Ielo, 5 Maretto, 6 San Lazzaro, 7 Torcio, 8 Dei Natali, 9 Capo Le
Mandre, 10 Matese lakes springs, 11 Letino lake springs, 12 Capo dacqua; 5) village;
6)mountain peak; 7) elevation (m a.s.l.); 8) endorheic area; 9) Torano river gauge; 10)
sinkhole: a-Brecce, b-Scennerato.
232 Karst without Boundaries
The long-term spring discharge measurements and the relation to climate vari-
able have been analysed by Fiorillo & Guadagno (2012), and the hydraulic aquifer
behaviour during droughts has been described by Fiorillo (2009) and Fiorillo etal.
(2012).
in the Volturno plain (Capriati plant hydropower), with a fall of 654 m. Lago Gallo
is also used as pump/storage reservoir.
Fiorillo etal. (2015) proposed an model to estimate the groundwater recharge at the
long-term scale, especially for large areas with strong morphological irregularities,
and not completely covered by hydrological monitoring. Based on long-term annual
mean data, the afflux (that is the total amount of meteoric precipitation), runoff, and
recharge are computed in GIS environment at yearly intervals, to estimate the recharge
and the runoff coefficient, for both the open areas and the endorheic areas. After esti-
mating the annual mean recharge, daily recharge has been estimated. Starting from
the beginning of the hydrological year, when field capacity has been reached, the daily
rainfall is split in (daily) recharge and (daily) runoff.
The annual model provides the long term estimation of the recharge. It is based
on GIS analyses, where the afflux, evapotranspiration and effective rainfall on the
catchment are estimated. The spatial distribution of the rainfall allows estimation of
the total amount of the afflux, F, in a specific area, A, by:
n
P
( F )A = 1
n
A (16.1)
where n is the number of cells in the area A, and P is the annual mean rainfall assigned
to each cell in as a function of its ground elevation, and deducted from the regression
of available rain gauges, located at different elevations.
234 Karst without Boundaries
If the actual evapotranspiration, AET, is subtracted from the rainfall, the effective
afflux, Feff, in a specific area, A, is:
n
P eff
(F )
eff A = 1
n
A (16.2)
In endorheic areas, AE, as the runoff cannot leave the catchment, the recharge
amount, R, can be considered equal to the effective afflux, Feff:
(R)A E
( )
= Feff
AE
(16.3)
If the water abstraction occurs inside the endorheic areas, the (net) recharge can
be considered to be reduced:
(R)A E
(
= Feff - QP ) AE
(16.4)
where QP is the annual water amount abstracted from the endorheic areas. The value
of QP has to be known in the case of abstraction from wells for human and agricultural
activity; the amount of water subtracted for hydroelectrical purpose was estimated.
In the open areas the recharge amount, (R)Ao, can be estimated assuming that all
the groundwater flow feeds the spring discharges, Qs, and no-flow boundaries occur
towards the argillaceous, terrigenous and flysch sequences (impervious terrains). With
this assumption, the total discharge, Qs, from springs is:
(R)Ao = Qs - (Feff - QP ) AE
(16.6)
(R)A (R)Ac
(CR )AE = 1; (CR )AO = ; (CR )Ac = (16.8)
O
To estimate the fraction in the spring discharge provided by each endorheic area,
the index, CS, was computed:
(F - QP )
(CS )A
eff AE
= (16.9)
E
QS
(CS )AO
= 1 - (CS )A (16.10)
E
The daily model is based on the balance of the water content of the soil the mantle
provided by (Fiorillo & Wilson, 2004), transforming the daily rainfall to water con-
tent, up to field capacity, and then to recharge and runoff.
In particular, the daily rainfall, Pi, is divided into several parts:
Pi = AET + Q + R + RO (16.11)
- Q = AET (16.12)
Pexc = R + RO (16.14)
which are the part of the rainfall which is free in the system (not transpired and not
retained as soil moisture). In this model the amount + R constitutes the infiltration.
In the endorheic areas, AE, as the runoff, RO, cannot leave the catchment, all the
excess rainfall, Pexc, can be considered as recharge:
In the open areas, AO, the model assumes that runoff occurs if rainfall excess, Pexc,
exceeds a specific threshold value, Tr, at a daily scale:
for Pexc > Tr , (Pexc )AO = (R)AO + (RO)AO, and (R)AO = Tr (16.17)
The threshold value, Tr, corresponds to the infiltration capacity at a daily scale
when the soil has reached field capacity. Tr is deducted in the simulation, for a specific
hydrological year, the ratio between the annual recharge amount in open areas, S(R)Ao,
and annual excess rainfall, SPexc (both computed summing daily values), is equal to
(CR)Ao computed at long-term annual scale (Equation 8). In particular, the threshold
can be estimated by considering several hydrological years characterised by annual
rainfall near the mean.
Some variables need to be fixed on the basis of in situ and laboratory tests, to
allow simulation of the evapotranspiration and recharge processes on daily scale to
be undertaken. The soil moisture capacity is needed, which depends on the thickness
of the soil involved in the evapotranspiration processes and on the range of the soil
moisture during the hydrological year (Qmax - Qmin).
In conclusion, the equations regarding annual and daily scale recharge, described
in Fiorillo etal. (2015), are computed analysing the morphological features of the
catchment area, which were divided in two main areas: the endorheic and open areas.
(RO)A O
= (RO)Ac = (Feff )AO - (R)Ao (16.18)
Tables16.1 and 16.2 show the results for the long term annual scale obtained for
each massif. The effective recharge coefficient computed for the open areas, (CR)Ao,
appears to have comparable values for the Cervialto (0.66) and Terminio (0.67) massifs.
The values of (CR)Ao = 0.66 has been used to estimate the recharge of the Matese massif,
where the total outflow from the massif is partially known (Fiorillo & Pagnozzi, 2015).
For the entire spring catchment, the effective recharge coefficient, (CR)Ac, is a func-
tion of the endorheic areas, and it is maximum for the Terminio massif, where the
Piana del Dragone has an important role in increase the recharge.
(mm/year)
N 800900 1.5001.600
9001.000 1.6001.700
W E
1.0001.100 1.7001.800
S 1.1001.200 1.8001.900
1.2001.300 1.9002.000
1.3001.400 2.0002.100
1.4001.500 2.1002.200
0 2 4 8 Kilometers
Figure16.4 Effective rainfall distribution for Terminio and Cervialto catchments (Fiorillo et al., 2015).
Endorheic areas are distinguished by yellow line, and relative items are described in
Table16.1.
N (mm/year)
400500
W E
500600
S 600700
700800
800900
9001.000
1.0001.100
1.1001.200
1.2001.300
1.3001.400
1.4001.500
1.5001.600
1.6001.700
1.7001.800
1.8001.900
1.9002.000
0 5 10 20 2.0002.100
Kilometers
2.1002.200
2.2002.300
Figure16.5 Effective rainfall distribution for the Matese massif (Fiorillo and Pagnozzi, 2015). Endorheic
areas are distinguished by yellow line, and relative items are described in Table16.2. The
red line splits the upwind (southern) and downwind (northern) zones.
Table16.1 Main hydrological parameters of Cervialto and Terminio catchments (modified from Fiorillo etal., 2015). F, afflux, Feff, effective afflux; RO,
runoff; QP, groundwater abstracted; CR, effective recharge coefficient; CR, total recharge coefficient; Cs, effective contribute to spring discharge.
Minimum
elevation F Feff R Qp RO
Item Area m (a.s.l.) km2 m3106 m3106 m3106 m3106 m3106 CR CR CS
1 Piano Laceno 1047 20.5 43.8 32.9 32.9 0.0 0.0 1 0.75 0.256
2 Piano Acernese 1168 3.3 7.4 5.8 5.8 0.0 0.0 1 0.78 0.045
3 Piano dei Vaccari 1164 1.4 3.0 2.3 2.3 0.0 0.0 1 0.77 0.018
CERVIALTO
4 Valle Rotonda 1156 1.1 2.3 1.8 1.8 0.0 0.0 1 0.78 0.014
5 Raia dellAcera 1246 0.7 1.5 1.2 1.2 0.0 0.0 1 0.80 0.009
Closed areas. AE 27.0 58.0 44 44.0 0.0 0.0 1 0.76 0.343
Open areas. AO 83 172.4 128.3 84.5 0.0 43.8 0.66 0.49 0.657
Springs catch.. AC 110 230.4 172.3 128.5 0.0 43.8 0.75 0.56 1.00
1 Piana del Dragone 668 55.1 103.9 71.6 65.3 6.3 0.0 0.91 0.63 0.383
2 AcquadellePietre 1061 4.3 8.9 6.7 6.7 0.0 0.0 1 0.75 0.039
3 Campolaspierto 1279 2.3 5.1 4 4.0 0.0 0.0 1 0.78 0.023
4 PianidIschia 1210 2.1 4.6 3.5 3.5 0.0 0.0 1 0.76 0.021
TERMINIO
5 Piano di Verteglia 1177 2.1 4.4 3.4 3.4 0.0 0.0 1 0.77 0.020
6 Piana di Cetola 752 1.4 2.6 1.8 1.8 0.0 0.0 1 0.69 0.011
7 PianaSantAgata 1047 1.3 2.6 2.1 2.1 0.0 0.0 1 0.81 0.012
Closed areas. AE 68.6 132.1 93.1 86.8 6.3 0.0 0.93 0.66 0.509
Open areas. AO 94.3 180.2 125.0 83.7 0.0 41.4 0.67 0.46 0.491
Springs catch. AC 162.9 312.3 218.1 176.8 6.3 41.4 0.81 0.57 1.00
Table16.2 Main hydrological parameters of Matese massif. F, afflux, Feff, effective afflux; R, recharge; QP, pumped water; RO, runoff; CR, effective recharge
coefficient; CR, total recharge coefficient; Cs, effective contribute to spring discharge.
Minimum
elevation F Feff R Qp RO
Item Area m (a.s.l.) km2 m3106 m3106 m3106 m3106 m3106 CR CR CS
1 Matese lake 1007 45.6 91.2 69.0 0 23.7 0 0.66 0.50 0.104
2 Letino lake 870 22.4 33.8 22.8 0 7.9 0 0.66 0.45 0.034
3 Gallo lake 822 14.2 20.3 13.0 0 4.4 0 0.66 0.42 0.020
4 Campitello Matese 1417 9.6 21.9 17.9 0 6.2 0 1 0.82 0.027
5 Stampata 1316 9.5 15.3 11.0 0 0 0 1 0.72 0.025
6 Campo Rotondo 1150 3.7 7.5 5.7 0 0 0 1 0.76 0.013
7 Cannamate 976 3.4 5.0 3.2 0 0 0 1 0.65 0.007
8 Campo Figliolo 1100 3.1 4.8 3.3 0 0 0 1 0.69 0.008
9 Campo Braca 1135 2.9 5.8 4.3 0 0 0 1 0.75 0.010
10 V. Campitello 1301 2.9 4.7 3.5 0 0 0 1 0.73 0.008
11 Vallecupa 1171 1.9 3.9 3.0 0 0 0 1 0.77 0.007
12 Pianellone 1290 1.6 3.5 2.7 0 0 0 1 0.79 0.006
13 Torricella 851 1.5 2.1 1.3 0 0 0 1 0.63 0.003
14 Tagliaferro 1096 0.6 1.2 0.9 0 0 0 1 0.74 0.002
15 Selva Piana 1286 0.9 1.1 0.8 0 0 0 1 0.78 0.002
Closedareas. AE 123.8 221.8 162.5 0 42.2 0 0.78 0.57 0.276
Open areas. AO 427.7 701.2 476.4 162.0 0 162.0 0.66 0.45 0.724
Springs catch. AC 551.5 923.0 638.9 162.0 42.2 162.0 0.69 0.48 1.000
240 Karst without Boundaries
The recharge coefficient, CR, assumes values always <1, as it is based on the total
amount of rainfall; considering the entire spring catchment, this value is 0.56 and
0.57 for Cervialto and Terminio, respectively, and provides a direct estimation of the
annual mean fraction of the rainfall which contributes to recharge. This value is 0.52
for the Matese massif (Table16.2).
The application of a daily scale model requires knowledge of some of the hydro-
logic parameters of the soil cover, which control the evapotranspiration processes and
retain an amount of water before recharge can occur. In particular, the soil undergoing
evapotranspiration depends also on the depth of plant roots and soil thickness. For
the Cervialto massif, the soil thickness undergoing evapotranspiration processes is
assumed H=50 cm, and the soil storage capacity is taken as m=120 mm (Fiorillo etal.,
2015). However, because the soil mantle thickness cannot be considered constant,
these values are an approximation.
The transformation of daily rainfall to daily recharge has been carried out in
Equation16.11, using a rain gauge characterised by annual mean rainfall similar to
the mean of the rainfall height on the karst massif.
Figure16.6 shows a graphic example of the simulation for the year 20112012,
where the total cumulative rainfall reaches a value of 1769 mm (Figure16.6A). In
Figure16.6B the daily rainfall is split according to Equation 11 for the daily scale
recharge model and during the initial period (SeptemberOctober), the daily rain-
fall is transformed into soil moisture increase, net of evapotranspiration. When field
capacity has been reached (Qmax=51%; Figure16.6b), daily rainfall is able to provide
100 2500
A) Pi=1768,6 mm
60 1500
40 1000
20 500
0 0
0 51
B)
vol. water content (%)
20 47
Pexc (mm)
43
40
39
60 runoff in open areas (or additional recharge in endorheic areas)
ET=545.0 mm 35
recharge
80 Pexc=1223.5 mm
threshold (22,3 mm) 31
volumetric water content
(R)Ao=877.2 mm
100 27
80 1.2
C) snowmelt
daily run off (mm)
60 RO=346.3 mm 1
40 0.8
20 0.6
0 0.4
1-Sep-
1-Oct-
1-Nov-
1-Dec-
1-Jan-
1-Feb-
1-Mar-
1-Apr-
1-May-
1-Jun-
1-Jul-
1-Aug-
2012
2012
2012
2012
2011
2012
2012
2011
2012
2012
2011
2011
Figure16.6 Hydrological characteristics of the 201112 hydrological year, Laceno rain gauge (1170m
a.s.l.) (modified from Fiorillo etal., 2015). A) daily and cumulative rainfall; B) volumetric
water content, Q, computed from daily hydrological balance (Qmin=27% and Qmax=51%)
and daily recharge (histogram); C) daily runoff (rainfall excess that exceeds 22,3 mm/day)
and river discharge measured in Caposele village.
Recharge processes of karst massifs in southern Italy 241
the daily rainfall excess amount, Pexc as well and the amount of recharge and runoff
an be estimated from Equations 16.15, 16.16 and 16.17. The threshold of 22.3 mm
(Figure16.6B) has been deducted from the ratio of total recharge amount in open
areas, S(R)Ao, and total excess rainfall, SPexc, equal to the effective recharge coef-
ficient, (CR)Ao=0,66 (Table16.1), for the hydrological year 197374 characterised
by the long term average annual rainfall. In open areas, rainfall excess that exceed
22.3mm is transformed into runoff.
Several critical aspects in evaluating the afflux in the mountain areas and the estima-
tion of the evapotranspiration were discussed by Fiorillo etal. (2015).
For the Matese massif, the runoff in the open areas is estimated at 34% of the
effective afflux, Feff (or 26% of the total afflux); this percentage is obtained using
results for the Cervialto massif. The percentage (in term of total afflux) used is lower
than values (3035%) found by Selmo (1930) who carried out the hydrological bal-
ance of Matese lake for hydro-electrical purpose during the 1920s, and higher (7.4%)
than Civita (1973) who assigned empirically the fraction of precipitation which infil-
trates into the aquifer.
The daily scale model has allowed the threshold for the daily recharge to be
assigned. This model considers a single point to estimate the recharge/runoff (1D
model), located approximately in the median zone of the Cervialto catchment. Thus,
the daily scale model provides an overestimation and underestimation of the recharge
in the lower and higher elevation zones, respectively. However, as it is calibrated
on the long-term annual scale recharge, it is a useful estimation of the mean daily
recharge in the catchment.
The total amount of recharge is a function of the rainfall distribution throughout
the hydrological year, and varies during wet or dry years. During wet years, the runoff
coefficient tends to increase and the recharge coefficient tends to decrease; opposite
hydrological behaviour occurs during dry years. The simulation shown in Figure16.6
has been replicated for several hydrological years, including dry and wet years, and
for different thresholds of daily recharge. Figures 16.7A and 16.7B show the depen-
dence of the recharge coefficient, (CR)Ao, from the maximum daily recharge (thresh-
old), in different hydrological condition: dry, normal and wet years. During wet years
the amount of runoff is higher than the mean, and (CR)Ao reaches minimum values; as
a consequence, the contribution from endorheic areas in recharging aquifers increases
during the wet years.
The daily estimation of recharge for the massifs has to be connected to the mean
of the entire catchment. Because of the wide range in ground elevation and the afflux
on the massif, the recharge estimation of a specific sector of the massif needs to be
calibrated as a function of the ground elevation and its exposure. However, similar
results found by the application of the daily scale recharge model for Matese and
Cervialto massifs allows this methodology to be transposed to other karst areas of
the Apennine, characterised by similar climate and morphology. The model could
also be used for other karst area world-wide, but needs a different hydrological
setting to consider the local climate and geomorphologic context, soil cover and
boundary flux.
1
0.75 0,77
0,66
(CR) Ao
0,55
0.5 dry year (197475)
22,3
a)
0
0 5 10 15 20 25 30 35 40
maximum daily recharge, mm
0,81
0.75
0,66
0,58
(CR) Ao
0.25
b)
0
0 5 10 15 20 25 30 35 40
maximum daily recharge, mm
Figure16.7 Recharge coefficient for open areas, (CR)Ao, function of threshold of daily recharge, for
different hydrological years; a) Cervialto (modified from Fiorillo etal., 2015); b) Matese
(modified from Fiorillo and Pagnozzi, 2015).
Recharge processes of karst massifs in southern Italy 243
REFERENCES
Bonacci O., Magdalenic A. (1993) The catchment area of the Sv. Ivan Karst spring in Istria
(Croatia). Ground Water 31(5), 767773.
Bonacci O. (2001) Monthly and annual effective infiltration coefficient in Dinaric karst: exam-
ple of the Gradole karst spring catchment. Hydrological Sciences Journal 46(2), 287300.
Celico P, Civita M. (1976) Sulla tettonica del massiccio del Cervialto (Campania) e le implica-
zioni idrogeologiche ad essa connesse. Bollettino Societ Naturalisti di Napoli 85, 555580.
Celico P, Magnano F, Monaco L. (1982) Prove di colorazione del massiccio carbonatico del
monte Terminio-monte Tuoro. Notiziario Club Alpino Italiano, Napoli 46(1), 7379.
Civita M. (1969) Idrogeologia del massiccio del Terminio-Tuoro (Campania). Memorie e Note
dellIstituto di Geologia Applicata dellUniversit di Napoli 11, 5102.
Civita M. (1973) Linfiltrazione potenziale media annua nel massiccio carbonatico del
Matese (Italia meridionale). Atti 2 International Conference Acque Sotterranee, Palermo,
129142 pp.
Drogue, C. (1971) Coefficient dinfiltration ou infiltration eficace, sur le roches calcaires
Actes du Colloque dHydrologie en Pays Calcaire, Besancon, 15, 121130.
Fiorillo F., Wilson R.C. (2004) Rainfall induced debris flows in pyroclastic deposits, Campania
(southern Italy). Engineering Geology 75(34), 263289.
Fiorillo F., Esposito L., Guadagno F.M. (2007) Analyses and forecast of the water resource in an
ultra-centenarian spring discharge series from Serino (Southern Italy). Journal of Hydrology
36, 125138.
Fiorillo F. (2009) Spring hydrographs as indicators of droughts in a karst environment. Journal
of Hydrology 373, 290301.
Fiorillo F. (2011) The Role of the Evapotranspiration in the Aquifer Recharge Processes of
Mediterranean Areas In Evapotranspiration From Measurements to Agricultural and
Environmental Applications, InTech, October 2011, Rijeka (Croatia), 373388 pp.
Fiorillo F., Guadagno F.M. (2012) Long karst spring discharge time series and droughts occur-
rence in Southern Italy. Environmental Earth Sciences 65(8), 22732283.
Fiorillo F., Revellino P., Ventafridda G. (2012) Karst aquifer drainage during dry periods.
Journal of Cave and Karst Studies 74(2), 148156.
Fiorillo F, Pagnozzi M. (2013) The role of endorheic areas on recharge processes of karst mas-
sifs. Internationl Sympsium on Hierarchical Flow Systems in Karst Regions, September 2013,
Budapest, Abstract volume, p. 65. ISBN 978-963-284-369-8
Fiorillo F., Pagnozzi M., Ventafridda G. (2015) A model to simulate recharge processes of karst
massifs. Hydrological Processes 29, 23012314.
Fiorillo F., Pagnozzi M (2015) Recharge processes of Matese karst massif (southern Italy).
Environmental Earth Sciences, DOI 10.1007/s12665-015-4678-y
ISPRA (2015) Geological Map of Italy, 1:50.000 Scale. Istituto Superiore per la Protezione e
la Ricerca Ambientale (ISPRA) Ministry of Environment: Rome. http://www.apat.gov.it/
Media/carg/
Parotto M, Praturlon A. (2004) The southern Apennine arc. In Special Volume of the Italian
Geological Society, Crescenti U, DOffizi S, Merlino S, Sacchi L (eds). Proceedings of the 32nd
International Geological Congress, Di Virgilio, Chieti, Florence; 3458.
Selmo L. (1930) Note idrologiche sul lago del Matese. LEnergia Elettrica, 7, 190199.
Thornthwaite, C.W. (1948) An approach towards a rational classification of climate.
Geographical Review 38, 5594.
Turc L. (1954) Le bilan deau des sols. Relations entre les precipitations, levaporation et
lecoulement. Annales Agronomiques, 5, 424.
White WB. (2002) Karst hydrology: recent developments and open questions. Engineering
Geology 65, 85105.
This page intentionally left blank
Chapter 17
ABSTRACT
Epikarst in South West China is important to the hydrological regime and consequent environ-
mental problems in the area. A slope with bare epikarst in the Fengcong area near Guilin has
been studied to understand the hydrological function of the surface layer. A physical model is
set up based on data from multiple methods of field observation, tracer tests and monitoring.
The water budget is used for comparing different hydrological links such as canopy intercep-
tion, soil moisture deficit, runoff generation and leakage/infiltration in the slope. Although in
the top layer of the slope, epikarst zone and soil infiltration dominates, but surface runoff still
happens during extreme storms and changes into mostly vadose flow within a very short dis-
tance. Results from this study show that the drainage area of a vadose stream varies depending
on climate. In the monsoon period it is about 491m2, which is larger than that in typhoon
period. However the average rainfall interception depth, which is the threshold for flow gener-
ation in both periods, is about 16mm. The threshold derives mainly from the canopy intercep-
tion, soil moisture deficit, and epikarst storage.
17.1 INTRODUCTION
Groundwater is plentiful in the karst areas in South West China. By contrast, there is
a shortage of surface waters. The karst environment is usually dry and suffers from
drought every year (Yuan, 1991; Guo etal., 2013). Residents in the karst mountain
areas sometimes have to pump groundwater from deep wells or karst windows to
combat water shortage. However, the cost is usually high because of the great thick-
ness (1001000 m) of the vadose zone. It is usually difficult to find groundwater
within the complex karst hydrogeological setting. In practice, the drilling failure ratio
reaches 1 out of 3 wells.
Due to the lack of rivers and the high cost of groundwater extraction, small water
tanks have been traditionally used to harvest rainwater and runoff from hill slopes.
Since 1999, in keeping with the Chinese Western Development Policy, thousands of
new water tanks were built in the Fengcong areas. Although these small tanks are
246 Karst without Boundaries
popular, they are vulnerable to water quality degradation and dry out in extreme
droughts. To improve the supply, tanks could be replenished more frequently with
new water from the hill slopes.
Hill slopes in the karst are often selected as rainwater collection areas and supply
water for the tanks. To optimise management and to guarantee the quantity and qual-
ity of water in the tanks, the hydrological processes on the slopes need to be studied.
Hill slopes in the Fengcong areas are usually convex, steep, with little soil and low
vegetation. Epikarst plays a key role in runoff generation and infiltration. The epi-
karst not only increases the water storage capacity in the surface layer (Williams,
1983), but also provides paths for rapid recharge. The morphology of the epikarst
zone exerts a significant impact on the hydrological process of each slope. This study
provides a closer insight into the epikarst and vadose zone by monitoring vadose
streams in a cave, and analyses the water balance by selecting a distinctive slope with
a developed cave. The results enhance understanding of the environmental problems
in karst areas.
Yaji Karst Experimental Site is about 8 km away from Guilin City, Guangxi Zhuang
Autonomous Region in South West China. It is an experimental research site for karst
study that started in 1986 (Yuan et al., 1990). The site is located in the border region
of Fenglin and Fengcong, both of which are typical landforms in Guilin and famous
throughout the world (Sweeting, 1995).
The climate in Guilin is characterised as subtropical Asia Monsoon. It is hot and
wet in summer, cool and moist in winter, with an annual temperature of 19.2C and
an annual rainfall of 1935 m. The precipitation in Guilin outs in two distinct periods.
One rainy season occurs between winter and spring, when the sun moves towards the
north and the moist winds from the ocean are strongest. It meets over Guilin with
the cool and dry wind from the north continent, and being well-matched in strength,
leads to a long rain period. Another rainfall period occurs during the summer and
autumn. It is often dry with several short rainstorms caused by typhoon.
A cave named Xiaoyan is developed in the vadose zone at the Yaji Experimental
Site. It is located in the Fengcong areas with the only entrance in the middle of a
west-facing slope. The Xiaoyan cave extends from west to east, with a total length
of 100 m and a variable width of one to 20 m. It occurs in massive limestone of
Devonian age and is controlled by fractures. The shape of the cave is wide in the
middle, but narrow in the tail (Figure17.1).
There is vadose water drainage from the ceiling at three places in the cave, located
at the entrance and in the middle of the cave. The first one XY1, located at the
entrance, is 4 m above the floor. The second one XY3, is 2 m high and flows along a
stalagmite from the top to the floor. The third one XY5, slumps in a hole at the ceiling
20m high, and then flows quickly in a film, forming a 10 m long flowstone. At the
end of the flowstone the stream becomes free falling. The thickness of the ceiling rock
is 40m.
The slope where the cave developed is 175m a.s.l., being 152 m a.s.l. at the foot
and 327 m a.s.l. at the topographic divides. The cave entrance altitude is 197ma.s.l.,
Water balance analysis of a vadose stream to discern hillslope hydrology 247
E
Cave
295
Cliff
285
275
265
255
245
235
235
245
225 185
215 205 195 0 20 m
Figure17.1 The slope with cave (shaded area) and survey points 1 soil moisture monitoring; 2 lysim-
eter; 3 surface flow I; 4 surface flow II; 5 surface flow III; 6 tracer injection I; 7 tracer
injection II; 8 tracer injection III; 9 cave stream XY5; 10 cave stream XY3; 11 cave stream
XY1; 12 cave entrance.
while the groundwater level is 150ma.s.l.. The vegetation on the rocky slope is shrub
dominated by Loropetalum chinense and Bauhinia championii (Benth.) Benth. Its
root grow in fractures and small holes. Soil fill in the open fractures is clayey. On the
slope, there are many cliffs and small ledges between the cliffs. The cliffs are shaped
by weathering and are controlled by vertical fractures. The cliffs range from 1 to 5m
high, with some conduits and flowstone developed, indicating that they play a part in
the flow discharge in the epikarst zone.
The terrain of the slope and the cave were mapped in order to show the surface karst
features and the projection of the cave on the slope. A 3 m long trench was dug in
a small platform in front of a cliff to take samples for assessing the soil physical
properties in the profile, including soil depth, structure, color and density. A mini-
permeameter was used to test the permeability of the soil. These tests were repeated
at three randomly selected points. There was no perennial flow at the surface. Short-
duration runoff in the slope was determined by field observation during storm events.
Hydrological monitoring of the surface flow was carried out several times.
248 Karst without Boundaries
was in the XY1 vadose stream under the injection point. One sample was taken after
each subsequent rainfall event.
Rainfall was measured at a raingauge station in the Yaji Experimental Site.
Precision of the measurements was about 0.2mm. Intercepted rainwater from the
canopy was obtained by the rain gauges. Three rain gauges were installed under the
canopy. The rainfall amount was compared with the gauge station in Yaji. The differ-
ence was the amount of rainwater intercepted by the canopy.
A simple micro-lysimeter was used to measure evaporation in the soil. The
lysimeter was made of PVC cylinders that had a diameter of 10cm and lengths of
5, 15, and 50cm. A slightly disturbed soil pillar was put in the cylinders, and the
nylon net was fixed at one of the ends. After weighing at 8a.m., the cylinder was
put back in the pit where the soil pillar was taken. The cylinders were weighed
again at 7p.m. in the same day. The weight difference represents evaporation in
the day and night. The test was performed continuously for seven days during July
and August, 2012.
In the second tracer test, the injection point was close to the first one. However, in
this test the tracer was only detected in XY5. Several samples taken from the surface
flow and drip water in the cave contained no tracer. The breakthrough curve of the
tracer in XY5 only lasted for four months. It recovered after being washed by several
storms in the rainy season, suggesting that the tracer cannot enter the epikarst at this
time. But, there is a negative peak in the concentration due to the dilution of baseflow
in the epikarst by surface flow leakage. It may be concluded that the tracer injected
in the first test was still stored in the vadose zone. However, it is still unclear how the
tracer enters into the epikarst in the first tracer test and where the open joint and shaft
is connected with the cave in the slope surface. The topography in the slope is not the
only factor which controls the flow. The subsurface space formed by karst processes
can also strongly affect the flow regime.
a stalagmite. It has a medium delayed time, and disappears for several months during
the year. The discharge of the stream has only been measured roughly because it is
spread into many segments by the stalagmite. The measured flow discharge ranges
from 01 ml/s.
The three vadose streams provide an opportunity to understand the behaviour of
vadose flow under the slope. Although they are all related with surface flow and epi-
karst, the hydraulic relationships and flow regime have a slight difference. Probably,
in the perched vadose zone system, the XY3 is in the downstream of XY5, while the
XY1 is isolated. It may imply that flow generated on the slope is concentrated at
different levels and moves along independent pathways. The vertical vadose flow is
transformed from lateral flow and surface flow and breaks the integrity of flow on
the slope.
17.5 DISCUSSION
slope ridge
Atmosphere
1 2 3
travertine Surface
4
discharge
(only in storm)
soil
tracer injection 1,2
sinkhole 5 6
Figure17.2 Schematic diagrams of the water cycle in the bare karst slope: left: slope profile; right:
diagram of flow.
Water balance analysis of a vadose stream to discern hillslope hydrology 253
soil and epikarst, where the roots absorb water. Epikarst water enters into the vadose
zone by flows and seepage in the paths 8 and 9, while some of them water returns to
the surface through epikarst springs, and some returns to the atmosphere by evapo-
ration from the epikarst. When the flow lines from paths 1, 5 and 8 in the karst slope
are overwhelmed, drought or flood will occur. The relative weakness of flow paths 3,
6 and 9 makes vegetation grow slowly and the ecosystem becomes fragile.
R=L(P-E),
R = Q+S
The amount of Q entering the epikarst will be reduced by infiltration in voids such
as the conduit wall and by algae and litter. The porosity of the rock surface enhanced
wherever lichens are present (Cao et al., 1998). Q should needs to be corrected as,
Q=Q0-hq
soil
runoff rock
soil
rock fracture interflow fracture
Figure17.3 Model of flow paths in the slope land surface, the plane and in section.
254 Karst without Boundaries
Where Q0, h, and q are the initial value of Q, the distance, and water absorption
by conduit per unit length, respectively.
Interflow in the slope IF by path 6 can be expressed as:
IF=S-D
In the equation D is the soil water deficit. It can be calculated by the equation:
D=qf-q1
(P-E)(1-L)=q1-q0
P-E=Q0+IF+qf-q0+h q
The threshold value for runoff on the slope will satisfy the water deficiency in the
soil and epikarst, so it can be defined as:
T=qf-q0+h q
M=Q+IF
P-E=M+T
M=P-E-T
For a vadose stream with a drainage area of A, the water balance equation
changes into:
M=A(P-E-T)
45
40
Monsoon period
35 Typhoon period
20
15
10
0
0 0.02 0.04 0.06 0.08 0.1 0.12
Rainfall (m)
Figure17.4 Relationship of rainfall and vadose stream XY5 discharge.The fitted results can be used to
determine drainage area and runoff decay.
256 Karst without Boundaries
bigger than that during storms. The statistics also shows that the discharge of XY5
does not change until rainfall reaches 50 mm in average.
The vegetation will take up water from the soil and the epikarst. A test of evapo-
transpiration was carried out in the karst environment in Guilin. The result shows
that a tree will take up 600 mm water in a normal year (Huang et al., 2013). The
vegetation in the slope usually suffers from drought, and it is not because of rain-
fall shortage but because of insufficient storage in the slope. Runoff feeding into the
vadose stream has in a large proportion compared to the total discharge (about 70%
of flow) in the vadose stream in one year. This provides a good reason why the karst
environment is so easy to dry out and to flood.
Among the components of the water balance for a vadose stream, the initial water
content and soil mass are related to the threshold values. In bare karst the soil stored
in fractures has a critical role for preferential recharge. After soil erosion occurs the
flow may generate more easily and quickly and in large amounts.
17.6 CONCLUSIONS
A hillslope with a cave developed below provides a good opportunity to examine the
flow pathways in the vadose zone. Both surface and subsurface flow can be observed
and monitored, which provide a basis for water budget analysis especially in a narrow
layer close to the slope surface. The hydrological process, such as canopy interception,
soil evaporation and soil moisture can be monitored and evaluated by various meth-
ods. Through establishing a water balance equation, a linear relationship between
rainfall and discharge of a vadose stream can be determined. Two important parame-
ters, the drainage area and the threshold value for flow generation are obtained from
the correlation analysis. The results show that a vadose stream contains preferential
recharge which is related to the drainage area of the slope surface.
The hydrological functions of a typical karst rocky slope comprise a large pro-
portion of quick flow, a large ratio of subsurface flow and a small ratio of surface
runoff. The soil water and seepage in the rock is not large and changes quickly. The
hydrological characters of the hill slope make the karst surface environment dry and
fragile. Water resources in the slope are not small but they are difficult to use due to
rapid drainage to the vadose zone.
The high proportion of vadose flow in the hydrological balance of the slope is
consistent with the geomorphologic character of mountain areas. The water balance
model for the slope in the bare karst areas provides a significant understanding of the
environmental problems in karst areas.
ACKNOWLEDGEMENTS
Financial support was provided by the Chinese National Natural Science Foundation
(41102161, 41172231, and 41472239), Key Project of Guangxi Natural Science
Foundation (2013GXNSFDA019024) and the Chinese Geological Survey project
(1212010916063). We would like to thank the reviewers who read the first draft of
this paper for their constructive comments.
Water balance analysis of a vadose stream to discern hillslope hydrology 257
REFERENCES
Ford D., Williams P. (2007) Karst hydrogeology and geomorphology. Wiley, Chichester.
Guo F., Jiang G.H., Yuan D.X., Polk J.S. (2013) Evolution of major environmental geolog-
ical problems in karst areas of Southwestern China. Environmental Earth Sciences 69(7),
24272435.
Jiang G.H., Guo F., Wu J.C. (2008) The threshold value of epikarst runoff in forest mountain
area. Environmental Geology 55, 8793.
Huang Y.Q., Li X.K., Zhang, Z.F. (2013) Seasonal changes in Cyclobalanopsis glauca transpi-
ration and canopy stomatal conductance and their dependence on subterranean water and
climatic factors in rocky karst terrain. Journal of Hydrology 402(12), 135143.
MWR (Ministry of Water Resources of Peoples Republic of China). 1993. Code for liquid flow
measurement in open channel.
Sweeting M.M. (1995) Karst in China: its Geomorphology and Environment. Springer-Verlag:
Berlin.
Williams P.W. (1983) The role of the subcutaneous zone in karst hydrology. Journal of
Hydrology 61, 4567.
Yuan D.X. (Editors) (1991) Karst in China. Geological Publishing House: Beijing.
Yuan D.X., Drogue C., Dai A. D. (1990) Hydrology of the karst aquifer at the experimental site
of Guilin in southern China. Journal of Hydrology 115, 285296.
Cao J. H, Wang F. X. (1998) Reform of carbonate rock subsurface by crustose lichens and its
environmental significance. Acta Geologica Sinica 72(1), 9499.
This page intentionally left blank
Chapter 18
Hydraulic behaviour of a
subthermal karst spring
Blederija spring, Eastern Serbia
Vladimir ivanovic , Veselin Dragiic , Igor Jemcov &
Neboja Atanackovic
University of Belgrade, Faculty of Mining and Geology,
Department of Hydrogeology, Belgrade, Serbia
ABSTRACT
Karst aquifers are characterised as a highly heterogeneous media which affects the behaviour
of the groundwater outflow regime. The Blederija subthermal karst spring is located on the
south-eastern part of the Miroc mountains. Complex quantitative analysis has been carried
out in order to obtain new insights into the hydraulic mechanism of the discharge regime.
This is characterised by frequent mixing of a normal and higher temperature waters. Binary
karst hydrogeological system and the location of ponor zones causes a specific variation of the
catchment area, which is reflected in the discharge regime properties of the karst aquifer. Two
different approaches of types of hydrograph separation were applied in order to identify the
interconnected flow system e.g. base flow and fast (direct) flow components. The first one was
based on the discharge rate and second on the water temperature. Both analyses gave very sim-
ilar results, and indicate a complex hydraulic mechanism, caused by a pressure pulse through
the karst hydrogeological system.
18.1 INTRODUCTION
Karst spring hydrographs are often used for characterisation of a karst aquifer.
Discharge analysis and correlation with the changes of climate parameters can be
used to understand the functioning of the karst system (Jemcov & Petric, 2010).
This is particularly important for gravitational springs where discharge variates sig-
nificantly and the influence of climate is pronounced. Deep siphonal springs have
a smaller discharge range and lower temperature variations. But, there are some
karst subthermal springs with significant changes in capacity and temperature which
result from mixing with colder karst groundwaters. In such circumstances, time
series regime analysis can help to understand the discharge mechanism as well as the
recharge conditions. This methodology has been applied at the Blederija subthermal
karst aquifer.
Blederija is a karst spring which drains the Miroc karst massif in eastern Serbia.
The karst aquifer is formed in fractured and highly-karstified massive Upper Jurassic
limestones (Stevanovic et al., 1996). A significant portion of the spring catchment
area is made up of low permeable Cretaceous clastic rocks, enabling the formation
of a network of surface streams, which sink as they pass through the karstic area
(Figure18.1). At times of high-water flows, most of the surface streams sink via the
ponor Cvetanovac. At the discharge zone, a cold spring and a subthermal spring (called
260 Karst without Boundaries
0 0.5 1 2
Spring Geology units Km
Springs catchment al Sand
Ponor Ng Sand, Clay
Groundwater flow direction K22,3 Siltstone, limestone
Temporal surface flow K15 Sandstone
Aquifer types K13,4 Limestone
Karst aquifers K11,2 Siltstone, Limestone
Fissured aquifers K11+2 Sandstone, Claystone
Intergrannular aquifers J3 Limestones
Low permeable rocks J2 Sandy limestones
Gp, Gneiss
mixed spring in this paper due to the results of regime monitoring) are located 10 m
apart. The average discharge of the Blederija springs is 280 l/s (Prohaska et al., 2001).
Monitoring was carried out for three years in order to understand the recharge and
discharge mechanism of Blederija spring (Figure18.2). Temperature was measured
at three locations (Blederija mixed spring, Blederija cold spring and Blederija stream
10m below the confluence of the mixed and cold spring), surface water level and
flow of the Blederija stream were measured, weather measurements: precipitation,
air temperature and wind-speed were also recorded. The measurements period was
March 2011 to April 2014.
The water temperature was measured every 5 days with digital and mercury ther-
mometers. In order to assess the capacity of Blederija stream, a water level gauge was
installed and read daily. Flow measurements were also conducted several times in
order to obtain a useable rating curve. Flow measurements represent a total flow of
both the cold and mixed springs, and only occasional measurements of subthermal
springs were performed during low-flow period.
In the vicinity of the spring, a meteorological station recording precipitation,
air temperature and wind speed and direction measurement was also installed.
Precipitation data were also collected from the Miroc rainfall station located near the
Hydraulic behaviour of a subthermal karst spring Blederija spring, Eastern Serbia 261
N
Tcold
Qcold* Tmix
Qmix*
Tsum
Legend:
Qsum Mixed spring
Cold spring
T Temp. measurement
Qsum Flow measurement
Qmix*, Qcold* Calculated flow
upper parts of the springs catchment. Data were also collected from the Negotin and
Crni Vrh meteorological stations where long-term digital climate event monitoring is
undertaken.
Beside the regime monitoring, field explorations were carried out, including tracer
test. Temperature and flow data were used for hydrograph separation. Time series
analysis was used to understand hydraulic behaviour of the Blederija karst system.
Together the data fed into the conceptual model of the aquifer.
The hydrograph separation method was applied in order to identify the base flow
and the fast flow components. There are numerous methods available to separate
baseflow from measured stream /spring flow hydrographs, but they are all based on
two concepts (Kyoung et al., 2010). One is based on a simple automated smoothing
and separation rule (Sloto & Crouse, 1996), and the second on a digital filtering
method that provides consistent results (Nathan & McMahon, 1990). A recursive
digital filter was used Eckhardt filter separates baseflow and fast flow components
(Eckhardt, 2005):
(1 - BFImax ) + bt -1 (1 - a)BFImax Qt
bt = (18.1)
1 - BFImax
Where bt is the filtered baseflow at the t step; bt-1 is the filtered base flow index at
the t-1 step; BFImax is the maximum value of long-term ratio of baseflow to total flow;
is the filter parameter; and Qt is the total flow at t step.
262 Karst without Boundaries
All data were processed using the BFI+3 module of HydroOffice 2015 (Gregor,
2010). To avoid uncertainties of estimation of the main parameter for the BFImax reces-
sion curve analysis, occasional flow measurements of the subthermal karstic spring
were used. Based on the 5 day data level, BFImax 0.45 (for the subthermal spring) and
parameter of 0.8 were estimated (Figure18.3).
In the next step, a different method of hydrograph separation was performed.
This method is based on the water temperature, because direct measuring of the
spring discharge was not possible due to the narrowness of the spring discharge area.
Spring discharges were calculated according to heat and mass continuum equations
(Kobayashi, 1985):
where Qsum is the Blederija stream flow, Qcold is the Blederija cold spring discharge,
Qmix is the Blederija mixed spring discharge, m is water mass, c is specific heat and Tis
water temperature.
Combining the upper two equations, hydrograph separation was obtained
(Figure18.4) based on the following relation:
Qmix=Qsum(Tsum-Tcold)/(Tmix-Tcold)(18.4)
3000
2500
2000
Capacity (I/S)
Qsum
1500 Qbase
1000
500
0
22.03.2011
05.06.2011
19.08.2011
02.11.2011
16.01.2012
31.03.2012
14.06.2012
28.08.2012
11.11.2012
25.01.2013
26.03.2013
09.06.2013
23.08.2013
06.11.2013
20.01.2014
05.04.2014
3000 20
18
2500
16
14
2000
Temperature (C)
12
Capacity (I/S)
1500 10
Qsum
8
Qmix
1000
Tmix 6
Tcold
4
500 Tsum
2
0 0
03.2011
05.2011
07.2011
09.2011
11.2011
01.2012
03.2012
05.2012
07.2012
09.2012
11.2012
01.2013
03.2013
05.2013
07.2013
09.2013
11.2013
01.2014
03.2014
Figure18.4 Hydrograph separation based on thermal equilibrium.
Table18.1 Descriptive statistics of summary discharge rate of Blederija spring and separated flow
components (l/s).
hydrograph components Qbase and Qmix. This confirms the applicability of the method
based on water temperature. Some discrepancy of the data between separated com-
ponents, Qbase, and Qmix, reflects the different methods applied and their parameter
calibration.
The temperature measurements have shown some unexpected results. The tem-
perature of the Blederija mixed spring in dry periods reached as high as 18.5C. In
the periods of snow melting, the mixed spring capacity significantly increases and
temperature decreases by almost 10oC. In previous studies (Dragisic et al., 1992;
Stevanovic et al., 1996), this phenomenon was explained by simple mixture with the
groundwater from Blederija cold spring. However, in the highest water flow periods,
the temperature of the mixed spring falls below the temperature of Blederija cold
spring (Tmix=8.1C, Tcold=9.7C). This suggests that the main cause of the lowering
of the water temperature of the Blederija mixed spring might not be the mixture
with the water from the cold spring. The temperature decrease could be due to mix-
ture with groundwater coming from the eastern part of the catchment, where large
264 Karst without Boundaries
portion of surface water recharges the karst aquifer via the Cvetanovac ponor. Dye
test conducted in December 2013 has confirmed this. The tracer (2 kg of uranine)
was injected in the ponor Cvetanovac with a river flow rate of 4 l/s (Figure18.5).
The tracer was detected 2.5 days later, but only at the Blederija mixed spring, and
not at the cold spring. During that time, the karst groundwater carrying the tracer
travelled 3060 m along a hydraulic gradient of 0.044 with an average flow velocity
of 1.4 cm/s.
In order to properly analyse the recharge-discharge relationship, the input com-
ponent of the karst system (precipitation) was converted to effective precipitation.
In order to separate the various processes in the air, vegetation and the soil from the
processes within the karst aquifer system, the interception on vegetation cover, and of
snow and snowmelt were assessed to define the effective precipitation, or the quantity
of precipitation that reached the ground (Jemcov & Petric, 2009).
One of the methods to study karst aquifer behaviour is time series analysis of the
recharge and discharge data as functions of the karst hydrogeological system (Jemcov
& Petric, 2010). The analysis has been carried out as complementary method to the
other field methods of karst groundwater exploration such as the dye test.
The univariate analysis characterises the individual structure of time. The auto-
correlograms of all of three functions (Figure18.6.) exceeds the confidence limits for
approximately 1315 time lag (~6575 days), which implies that the system is well
structured, and storage is significant.
The slope of autocorrelogram Qcold initially drops quickly for less than 15 days,
and afterwards remains on the same trend. This bimodal behaviour indicates the
Figure18.5 Dye test verified connection of the Cvetanovac ponor and Blederija mixed spring.
Hydraulic behaviour of a subthermal karst spring Blederija spring, Eastern Serbia 265
duality of the outflow regime. The same occurs in autocorrelograms Qsum and Qmix
(but less pronounced), and indicates a zone of mixing of water with normal and
subthermal components.
The bivariate analysis considers the transformation of the input to the output
signal. The cross-correlation function (CCF) of the Blederija source for all consid-
ered components (Figure18.7.) shows non-symmetrical behaviour and a high level
of influence of effective precipitation on outflow components, particularly within the
initial time lag (5 day). This suggests that all of the analysed components recharge
from the binary karst system. A function for all three outflow components becomes
insignificant after 55 days, and thereafter it exceeds the level of significance because
0.7
0.6 Qsum
Qhot
0.5
Qcold
0.4
Is
0.3
r(k)
0.2
0.1
0.0
0 5 10 15 25 30 35 40 45 50
0.1
0.2
k=5 day
Figure18.6 Autocorrelation function of discharge components of the Blederija source. Legend: Qsum
Blederija stream flow, Qcold Blederija cold spring Qmix Blederija mixed spring.
0.5
QsumPgi (swe)
0.4
QhotPgi (swe)
0.2
k=5 day
0.0
45 15 10 5 0 5 10 15 20 25 30 35 40 45 50
0.1
0.2
Figure18.7 Cross-correlation function of effective precipitation and flow components of the Blederija
karst source. Legend: Pgi (swe) effective precipitation transformed from the measured
values corrected for wind influence and interception by vegetation, and finally melted snow.
266 Karst without Boundaries
of the influence of the Qmix components, and the hydraulic behaviour of the out-
flow components. Constant variation of cross-correlation functions implies non-
homogenous karst hydrogeological system, with different responses in various parts
of the systems with frequent interchanges of normal and subthermal waters. A strong
influence of the effective precipitation on the Qmix component, similar to the Qcold
component, indicates a process of continuous interference of normal and subthermal
waters. Moreover, a common attenuation effect for the Qmix component is not quite
so obvious, and suggests possibilities of different development of the various levels of
the karst channel system (upper and lower).
Ponor
Cvetanovac Cold
spring
T=11.8C
Mixed
spring
T=18.5C
Ponor
Cvetanovac
snow melting T=4.1C Cold
spring
T=9.7C
Mixed
spring
T=8.1C
1 2 3 4 5 6 7 8 9
Figure18.8 Conceptual model of Blederija spring discharge during the low (up) and high (down)
water regime. (1) low permeable rocks, (2) karstified limestone (vadose zone), (3) karsti-
fied limestone (saturated zone with cold water), (4) karstified limestone (saturated zone
with thermal water), (5) groundwater level, (6) cold groundwater flow direction, (7) ther-
mal groundwater flow direction, (8) cold spring, (9) mixed spring.
Hydraulic behaviour of a subthermal karst spring Blederija spring, Eastern Serbia 267
Applied analysis and field exploration were complementary, and both were needed to
obtain the conceptual model. There is no doubt that water from the higher parts of
groundwater system has an effect on the decrease in temperature of the lower subther-
mal groundwater system. There are two independent upper groundwater karst systems,
coming from the north and the east, and only the water from the east system effects the
lowering of the water temperature of the subthermal spring. Therefore, the application
of the field techniques such as dye test along with other methods, e.g. isotope analy-
sis, are still essential to the exploration of the karst hydrogeological system. The com-
plex hydraulic relation between the cold and the subthermal groundwater in different
groundwater stages is presented schematically in the conceptual model (Figure18.8).
During the low water flow regime, the Cvetanovac River dries up and diffuse
infiltration occurs. During low water period, groundwater flows slowly, especially in
the deeper parts of the karst system. As a result of small spring capacities the water
temperatures become higher. The temperature of the mixed spring goes up to 18.5C.
In the periods of snow melt, punctual infiltration becomes pronounced, and intensive
recharge with very cold water (T=4.1C) occurs resulting in significant increase of
both spring flows, but also a significant decrease in water temperatures. Dye test
proves this phenomenon of temperature inversion for Blederija springs.
Time series analysis confirmed these results, especially the bimodal behaviour,
with strong diffuse infiltration in the binary karst hydrogeological system. This indi-
cates different responses in various parts of the systems with frequent interchanges of
normal with subthermal waters.
ACKNOWLEDGMENTS
This research was supported by the Ministry of Education, Science and Technological
Development (as a part of the Projects No. 43004 and No. 176022).
REFERENCES
Dragiic V., Stevanovic Z., Filipovic B. (1992) The occurrences of deep siphonal circulation
of karst aquifer of Miroc mountain (NE Serbia). Theoretical and Applied Karstology 5,
115120.
Eckhardt K. (2005) How to construct recursive digital filters for baseflow separation.
Hydrological Processes 19, 507515.
Gregor M. (2010) BFI+ 3.0 modul. HydroOffice 2015. URL: http://hydrooffice.org; last visited
10/10/2015.
Jemcov I., Petric M. (2009) Measured precipitation vs. effective infiltration and their influence
on the assessment of karst systems based on results of the time series analysis. Journal of
Hydrology 379(34), 304314.
Jemcov I., Petric M. (2010) Time Series Analysis, Modelling and Assessment of Optimal
Exploitation of the Nemanja Karst Springs, Serbia. Acta Carsologica 39(2), 187200.
Kobayashi D. (1985) Separation of the snowmelt hydrograph by stream temperatures. Journal
of Hydrology 76, 155162.
268 Karst without Boundaries
Kyoung J.L., Youn S.P., Jonggun K., Yong-Chul S., Nam W.K., Seong J.K., Ji-Hong J., Bernard
A. E. (2010) Development of genetic algorithm-based optimization module in WHAT system
for hydrograph analysis and model application. Computers & Geosciences 36(7), 936944.
Nathan R.J., McMahon, T.A. (1990) Evaluation of automated techniques for baseflow and
recession analysis. Water Resources Research 26(7), 14651473.
Prohaska S., Ristic V., Dragiic V. (2001) Groundwater budget and dynamical reserves estima-
tion of the Miroc karst massif (in Serbian) Proracun bilansa i dinamickih rezervi podzemnih
voda karstnog masiva Miroc. Vodoprivreda 33(189194), 3540.
Sloto R.A., Crouse M.Y. (1996) HYSEP: a computer program for stream flow hydrograph
separation and analysis. U.S. Geological Survey, Water-Resources Investigation Report
96-4040, Reston, VA, 46 pp.
Stevanovic Z., Dragiic V., Dokmanovic P., Mandic M. (1996) Hydrogeology of Miroc Karst
Massif, Eastern Serbia, Yugoslavia. Theoretical and Applied Karstology 9, 8995.
Chapter 19
ABSTRACT
The observed Plitvice Lakes Catchment area is located in the central part of the Dinaric karst,
in Croatia. It is the largest catchment area within the Plitvice Lakes National Park. Due to its
specific hydrogeological conditions, it can be divided into several subcatchmens. These were
determined through the synthesis of research results that includes hydrogeological character-
istics of the catchment area, hydrochemical and isotopic analysis of water from springs, and
hydrological analysis of the surface water system. The result is a division of the Plitvice Lakes
catchment into three main subcatchment areas: Matica, Plitvica and Jezera (Lakes).
19.1 INTRODUCTION
The Plitvice Lakes catchment area is approximately 152 km2. It is a part of the Korana
groundwater body (Biondic et al., 2013), which belongs to the Black Sea catchment
area, and is located in the central part of the Dinaric karst in Croatia. The Plitvice
Lakes catchment is the largest catchment within the Plitvice Lakes National Park
whose total surface area is 296.85 km. The wider Plitvice Lakes catchment area has
been designated a national park since 1949 due to its outstanding natural beauty. For
the same reason the Plitvice Lakes National Park has been included on the UNESCO
List of World Heritage Sites since 1979. This was the first area in Europe to be
included in this list due to the attractive phenomena created by water. It is the largest,
oldest and most visited Croatian national park.
The most prominent surface water occurrences within the Plitvice Lakes catch-
ment are cascading lakes of various sizes. The biodynamic process of predominant
tufa barrier growth created those lakes and nowadays there are 16 lakes, whose sur-
faces and forms are constantly changing over time (Figure19.1). The largest lakes
are Kozjak and Procansko. Beside the well-known lakes system, there are numerous
permanent karst springs in the Plitvice Lakes catchment, of which the most important
are the Crna Rijeka, Bijela Rijeka and Plitvica.
Nowadays the Plitvice Lakes National Park receives more than one million tour-
ists per year; urbanisation is encroaching upon the administrative boundaries of the
Park, causing the human impact to be pronounced in some parts. Therefore, for the
protection of the Plitvice Lakes National Park it is now no longer advisable, nor sus-
tainable, to restrict protection only to the narrow area around the lakes system but to
expand the protection zone out to the entire Plitvice Lakes catchment. It is necessary
270 Karst without Boundaries
SLO HUN
N
4456'0''N
BIH Slovenia
Hungary
BiH Serbia
Italy
ITA
tuk
Sar
N
Bijela
4450'30''N
rijek
anj
Su
a
vic
Ciganovac
t
Ljeskovac
Pli
Mat 37 m
ica Okrugljak
Batinovac
Veliko i Malo Korana
a Galovac 48.5 m
ek
Sastavci
Procansko Lake 24 m (583 m a.s.l.)
rij
Milanovac
(637 m a.s.l.)
na
Gradinsko
Cr
Novakovica Brod
4445'0''N
Gavanovac Kaluderovac
ica
Kozjak Lake
ec
Rj
(535 m a.s.l.)
Figure19.1 Location of the National Park Plitvice Lakes with position of the Plitvice Lakes catchment
area (left). Sketch of the lakes system (right).
Determination of the catchment area is the basis for all water balance calculations. It
enables the estimation of groundwater and surface water volumes, as well as the iden-
tification of possible sources, the direction and flow of water, and the transportation
of potential contamination in water.
Delineation of the subcatchment areas was carried out using three basic sets of
data: hydrogeological, hydrometeorological and hydrochemical. Hydrogeological
data represent the basic data that are needed to define the conceptual position of the
catchments and subcatchments. Hydrological and meteorological analyses were used
to quantify the presumed hydrogeological catchments, which were then confirmed by
iteration. Hydrochemical data were used for verification; these are mainly discrete
data that are influenced by conditions of spatial data in the catchment.
(Dimitrijevic, 1982; Herak, 1986, 1991; Pamic et al., 1998; Korbar, 2009) and it
can be noted that the evolution of the Dinarides has had a significant impact on
these processes. Hydrogeological characteristics and water permeability assessments
of layers (Figure19.2) were determined according to existing structural data and the
lithological composition of rocks.
The geological map for the area (Polak et al., 1967; Velic et al., 1970), the
Geological Map of Plitvicka jezera (Polak, 1969) and other relevant and available
data relating to the geology of investigated area were used (Bahun, 1978; Biondic &
Goatti, 1976; Biondic, 1982; Herak, 1962; etc.).
Data collected in the field, the position of sources and sinks, karst geomorpholog-
ical features as well as the results of tracing tests from the wider area were taken into
consideration during the hydrogeological analyses of the Plitvice Lakes catchment.
Research of the direction of groundwater flow in the Plitvice lakes has been carried
out on several occasions, either through systematic scientific research or as part of
other research. By the end of 2012 in the Plitvice Lakes catchment, 17 tracer tests for
groundwater flow were carried out (Biondic et al., 2008; Meaki, 2011).
SARTUK SPRING
LEGEND
permanent spring
4455'0''N
BIG WATERFALL
permanent watercourse
PLITVICA SPRING
temporary river
lake
water divide
main water divide
LIMAN DRAGA
SPRING border of the National Park
RJECICA
geol. boundary
VUKMIROVIC A
geol. boundary inferred
4450'0''N
SPRING
fault
fault inferred
reverse fault
BIJELA RIJEKA CRNA RIJEKA
SPRING SPRING "Kozjak" fault
low permeable clastics deposits
low permeable carbonates
LJESKOVAC
SPRING
medium permeable carbonates
high permeable carbonates
4445'0''N
0 2 4 6
km groundwater direction
The spatial distribution of rainfall was obtained from the available rainfall data
recorded in the reference period from 1961 to 1990 (DHMZ, 2009). Point data were
interpolated using the Spatial Analyst Tools module within the ArcGIS software pack-
age. The mean annual precipitation in the Plitvice Lakes catchment varies from 1224
to 1475 mm (see P in Table19.1), or approximately 1350 mm for the whole Plitvice
Lakes catchment. Isohyets highlight a north west to south east zone coinciding with
the dominant mountains in this area (Meaki, 2011).
Effective precipitation is a part of the total precipitation that can reach the satu-
rated zone and recharge the aquifer. Ultimately, this is the volume of spring discharge.
Usually it can be obtained through the application of various empirical formulae.
In the Plitvice Lakes catchment the empirical formula of ugaj (1995) was applied,
which is especially suitable for the Dinaric karst catchments in Croatia. Values varied
from approximately 790 mm in the southern part of the Plitvice Lakes catchment to
approximately 880 mm in the middle and northern part of the Plitvice Lakes catch-
ment (see Pe in Table19.1 and Figure19.2). The possible deviation of obtained result
is 18%.
The runoff coefficient (sometimes called effective infiltration coefficient) was
determined also from the precipitation data. This coefficient is defined as the ratio of
effective precipitation to total precipitation and it includes all processes that can occur
in some catchments during transformation of the total precipitation into effective
precipitation. Therefore, it is used very often in scientific and theoretical hydrological
analysis (Bonacci & Jelin, 1988; ugaj, 1995; Bonacci, 1999, 2001). The calculated
runoff coefficient for the whole Plitvice Lakes catchment is approximately 0.65 (see
c in Table19.1).
The analysis of the spatial distribution of air temperature includes the correlation
between altitude and normal annual air temperature at certain measuring stations in
the wider area for the reference period from 1961 to 1990. This is a linear regres-
sion equation, which represents the vertical temperature gradient, i.e. a drop in air
temperature of 0.5C per 100 m of altitude (Meaki, 2011). This is in line with a
similar approach that was made in previous research related to this part of Croatia
(Zaninovic et al., 2004).
In order to calculate evapotranspiration the empirical formula of Turc (1954)
and Coutagne (1954) were used. When using the Coutagne method it is essential
that the correct equation is chosen, in accordance with Coutagne parameter . In the
Plitvice Lakes catchment the Coutagne equation: precipitation 0.5/, was selected.
The resultant average values of evapotranspiration, based on both empirical formu-
lae, are in the range of 450 to 500 mm (see Et in Table19.1). These methods have
been used for a very long time for estimation of evapotranspiration in various areas,
including karst. As these are empirical formulae, they should be applied with caution
when used in a certain area. In the Plitvice Lakes catchment area several methods were
tested, and ultimately this two were selected as the most appropriate.
The average annual evaporation from the lake surfaces has also been calculated.
The calculation is based on an empirical relationship (Meyer, 1915) using measure-
ments from the climatological station located near Kozjak Lake. The value obtained
is 422 mm 19 mm (Meaki, 2011).
For the hydrologic analysis available data on discharges measured at water gaug-
ing stations within the Plitvice Lakes catchment (DHMZ, 2009) were used (see Qav.8108
Table19.1 Calculated hydrometeorological values for the Plitvice Lakes catchment (PLC) compared to measured flow rates (period from 1981 to2008).
% of P Pe Et 2
Qav.81-08
1
Description PLC Area (km )
2
(mm) (mm) (mm) c (Pe/P) Qcalc (m /s)
3
(m3/s) 3
g.s.
Bijela River with Ljeskovac 13 20 1224 774 449 0.63 0.50 0.448 2
Crna River 41 62 1245 791 454 0.64 1.55 1.36 1
Matica (direct) 1 2 1273 798 475 0.63 0.042 no data
Subcatchment 55 84 1247 788 459 0.63 2.09 2.09 3
MATICA
Upper lakes (direct catchment) 10 10 1376 893 483 0.65 0.276 no data
Suanj 3 5 1257 797 460 0.63 0.125 0.055 4
Rjecica River 8 12 1455 969 486 0.67 0.383 0.426 5
Matica inflow +2.09 2.09 3
evaporation from the lakes -0.025 no data
water intake from the Kozjak Lake -0.100 no data
Subcatchment 18 28 1363 886 477 0.65 2.75 2.65 6
JEZERA
Sartuk 7 10 1475 998 477 0.68 0.325 0.094 8
Plitvica spring 18 28 1374 919 455 0.67 0.807 no data
Plitvica River (direct catchment) 2 3 1452 959 493 0.66 0.077 no data
Subcatchment 27 41 1434 959 475 0.67 1.21 0.655 7
PLITVICA
PLC TOTAL 100 152 1348 878 470 0.65
1
For location of specific subcatchments, see Figure19.7.
2
Calculation is based on flow measurements at gauging stations.
3
For location of specific gauging station (g.s.), see Figure19.4.
274 Karst without Boundaries
in Table19.1). It should be noted that most of the gauging stations only started oper-
ating in the early 1980s, at a time when the majority of streams in the Croatian karst
region had sustained long dry period (ugaj, 1995). Therefore, to obtain a sufficiently
long time series of annual flow, analyses of homogeneity of time series were carried
out using Wilcoxson (1945) nonparametric test. The length of the hydrological series
is validated according to the error size of the variation coefficient, using the equation
by UNESCO (1982) and Kritsky-Menkel (1961).
Time series of the flow as well as identification of possible changes in the hydro-
logical regime were analysed using the Rescaled Adjusted Partial Sums (RAPS) method
(Garbrecht & Fernandez, 1994). Biondic (1999) and Bonacci & Andric (2008) car-
ried out similar analysis of hydrological changes in other karstified terrains using the
RAPS method.
The cyclicality of the hydrological regime in the Plitvice Lakes catchment
was analysed using RAPS (Figure19.3). In 50 years of hydrological observations
of the Plitvice Lakes catchment area, exchange of three wet and three dry peri-
ods were detected, i.e. every 16 years there is an extreme wet and dry period.
However, for now it cannot yet be seen a continuous trend of decline or increase
in the amount of water in the Plitvice Lakes catchment. It also should be noted
that the individual annual extremes seen in Figure19.3 do not need to fit into the
perceived frequency of the hydrological regime and that the calculation are based
on hydrological observation related to gauging station marked with number 9
(see Figure19.4).
The final analysis of the water balance shows that the Matica flow represents a
major contribution of water to lakes system, approximately 2.09 m3/s of water, more
than all the other inflows put together. From this amount of water, about 74% is from
Crna River and about 24% of water is from Bijela River (Figure19.4).
40
30
RAPS(Pl.Lakes) = 10sin( Tmy/96 /2) + 11
20
RAPS
10
10
20
1/1959
1/1962
1/1965
1/1968
1/1971
1/1974
1/1977
1/1980
1/1983
1/1986
1/1989
1/1992
1/1995
1/1998
1/2001
1/2004
1/2007
Remarks: value of the number; Tmy month and year for which the RAPS is
determined (in this case the initial time is January 1959)
Figure19.3 The RAPS analysis of mean monthly flow in the Plitvice Lakes catchment outflow, with
marked periodicity.
Delineation of the Plitvice Lakes karst catchment area, Croatia 275
uk
art
8S
Bije Q Plitvica spring (calc)
la ri
jeka 0.81 m3/s
j
Q Bijela River (calc) 4 Su an
Q Sartuk (calc)
0.50 m3/s
0.325 m3/s
ica
tv
Ljeskova 2 Plitvica Bridge
Pli
c Q
1 Mat (average '81'08)
ica 3 0.66 m3/s
a
6 7
k
2%
ije
Korana
r
Q Matica (calc)
na
Crna River
Upper lakes
Matica (direct catchment) (direct catchment)
Rjecica
9%
12% 4% Suanj
total 81%
evaporation
Matica 11%
68% 3% water capture
inflow
3% unknown
Kozjak Bridge
N
Plitvica Bridge
m 1 m3/s gauging station with corresponding number
0 1000
Figure19.4 Visual comparison of the calculated flow in the Plitvice Lakes catchment in the period
from 1981 to 2008.
Details and more results from testing the homogeneity of time series of annual
flow at water gauging stations in the basin lakes, the length of individual datasets, as
well as calculations of the error coefficient variation can be found in Meaki (2011).
Table19.2 The average concentrations of major ions at certain observation points (period from 2005
to 2007).
Bijela Rijeka spring (BR-I) 28 22 0.4 0.1 48 0.4 0.8 0.6 10.55
Crna Rijeka spring (CR-I) 36 14 0.4 0.1 48 0.5 0.7 0.5 9.08
Matica River (M-R) 33 17 0.3 0.1 48 0.3 0.6 0.4 9.46
Rjecica River (R-R) 26 24 0.2 0.1 49 0.2 0.4 0.3 11.16
Plitvica spring (PL-I) 32 18 0.2 0.1 48 0.6 0.4 0.8 9.78
Sartuk spring (SAR-I) 27 24 0.2 0.1 48 0.1 0.4 0.6 10.26
Ljeskovac spring (LES-I) 33 18 0.2 0.1 47 0.2 0.5 0.7 10.31
Rjecica (RJ-I) 27 24 0.1 0.1 49 0.0 0.4 0.0 11.22
PLC average based on all 29 21 0.3 0.1 49 0.4 0.5 0.4 9.73
data (Biondic et al., 2008)
For location of specific water sampling points, see Figure19.7.
1
6.0
Bijela rijeka spring
Crna rijeka spring
5.6
Plitvica spring
Rjec ica river
HCO3 (meq/l)
5.2
4.8
4.4
4.0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Ca /Mg
2+ 2+
Figure19.5 Comparison of the HCO3- concentration and the ratio of Ca2+ and Mg2+ in source waters
(period from 2005 to 2007).
Sampling and analysis of stable isotopes of water was carried out, which aimed
to complete the picture of the natural water system; the identification/verification
of the mean recharge altitude of springs, the determination of the mean residence
time of groundwater as well as understanding the processes of ionic changes in
water and evaporation. The Local Meteoric Water Line (LMWL) was also obtained
(Figure19.6). The content of stable isotopes 18O and D in precipitation was deter-
mined from a composite sample of rainwater for the period 20032007.
Delineation of the Plitvice Lakes karst catchment area, Croatia 277
66
Bijela rijeka spring
68 Matica River
main spring
LMWL Plitvice
zones Rjec ica River
D = 7.86 18O + 12.5
D ()
70 Plitvice spring
Plitvica River
74
10.8 10.7 10.6 10.5 10.4 10.3 10.2 10.1 10.0 9.9
18O ()
Figure19.6 Analysis of 18O/D values for all major and some minor sources and streams in the
Plitvice Lakes catchment area (period from 2005 to 2007).
1535'0"E 1540'0"E
SARI
River
N an
a
N
r
Ko
KR
LABEL DESCRIPTION
2.2 BRI Bijela Rijeka Spring
4455'0"N
2 PLITVICA RR
SARR Sartuk River
1.2 2.1 Plitvica spring
PROJ VUKI Vukmirovica Spring
2.2 Sartuk
1.3 2.3 Plitvica rijeka
3 JEZERA
3.1 Jezera (direct) LIMI
3.2 Rjecica MR RJI
1.1
3.3 Suanj
VUKI
LESI
BRI
4445'0"N
0 2 4 0 1 2 CRI
km km
Figure19.7 Main subcatchments in the Plitvice Lakes catchment (left). Name and position of major
water sampling points used for hydrochemical and isotopic analysis (right).
Figure19.8 Crna Rijeka spring (left) and Bijela Rijeka spring (right) (October 2005).
a series of minor karst springs, approximately 2 km before the junction with the Bijela
River. Higher parts of the basin consist of Upper Malm age carbonate sediments,
while the lower parts mainly consist of Doger age limestones. The total outflow from
the Bijela River and Ljeskovac Stream is measured on the Bijela River at the gauging
station of the same name (marked with number 2 at Figure19.4).
LEGEND
swallow hole
water divide
Hajdukovica mlin
Sastavci geol. boundary
Big Waterfall geol. boundary inferred
fault
fault inferred
4454'0"N
3 Subcatchment Jezera
Figure19.10 The Big Waterfall and Sastavci (source area of the Korana River).
ACKNOWLEDGEMENTS
A great deal of data was collected with the project Sustainable utilisation of water in
the pilot area Plitvice Lakes. The project was carried out within the Kompetenznetwerk
Wasserressourcen Gmbh, co-financed by the Plitvice Lakes National Park and the
Austrian government. The authors wish to thank all associates from the Joanneum
Research institute (Graz, Austria) and the Research Centre Ivo Pevalek (Plitvice
Lakes National Park, Croatia), who have made significant contributions and enabled
the successful completion of the project.
282 Karst without Boundaries
REFERENCES
Korbar T. (2009) Orogenic evolution of the External Dinarides in the NE Adriatic region:
a model constrained by tectonostratigraphy of Upper Cretaceous to Paleogene carbonates.
Earth-Science Reviews, 96, 296312.
Kritsky S.N., Menkel M.F. (1961) On the regularities of long term river flow fluctuations and on
the methods of estimation of lingering droughts. Hydrological Sciences Journal 6(2), 3440.
Kuhta M. (2010) Vodoistrani radovi u cilju zatite izvora Krbavica II. faza (in Croatian).
Archive of the HGI-CGS, Zagreb.
Meaki H. (2011) Model zatite krkih vodnih resursa na primjeru Nacionalnog parka Plitvicka
jezera (in Croatian). PhD thesis. Faculty of Mining, Geology and Petroleum Engineering,
University of Zagreb, Zagreb, 211 p.
Meyer A.F. (1915) Computing Run-Off from Rainfall and Other Physical Data. Trans. ASCE,
79, 10561224.
Pamic J., Guic I., Jelaska V. (1998) Geodynamic evolution of the Central Dinarides.
Tectonophysics 297, 251268.
Polak A. (1969) Geoloka karta Plitvickih jezera, M 1: 50 000. Kartografski laboratorij
Geodetskog fakulteta. Archive of the Faculty of Science, University of Zagreb, Zagreb.
Polak A., parica M., Crnko J., Juria M. (1967) Osnovna geoloka karta SFRJ M 1:100000,
list Bihac, L 33116. Archive of the HGI-CGS, Zagreb.
Stumm W., Morgan J.J. (1981) Aquatic Chemistry, 2nd ed. John Wiley and Sons, New York,
109 p.
Turc L. (1954) Calcul du bian de leau evaluation en fonction des precipitations et des tempera-
tures. IASH Symposium, Rome, 111(38), 188202.
UNESCO (1982) Methods of computation of low streamflows. Paris.
Velic I., Bahun S., Sokac B, Galovic I. (1970) Osnovna geoloka karta SFRJ, M 1:100000, list
Otocac, L 33115. Archive of the HGI-CGS, Zagreb.
Wilcoxon F. (1945) Individual comparisons by ranking methods. Biometrics, 1, 8083.
Zaninovic K., Srnec L., Percec Tadic M. (2004) Digitalna godinja temperaturna karta
Hrvatske. Hrvatski meteoroloki casopis, 39, 5158.
ugaj R. (1995) Regionalna hidroloka analiza u kru Hrvatske. Hrvatsko hidroloko drutvo,
Monografije br. 1, Zagreb, 139 p.
This page intentionally left blank
Part 4
ABSTRACT
Environmental indicators are a tool to measure results of human activities and their impacts
on nature and biodiversity. Water dependent ecosystems are under increasing pressure from
human activities and in karst they are exposed to greater potential hazard because of the high
vulnerability of karst aquifers. The aquifers of the Dinaric karst, one of the largest karst sys-
tems in the world, are abundant in water reserves; however there are several important poten-
tial threats to environmental and water quality values which have been identified through
environmental and socio-economic surveys conducted within the DIKTAS project. Five strate-
gic Environment Quality Targets are envisaged in this project: two (Groundwater sustainably
used; Water and soil quality controlled and improved) have a direct relationship with the
karst aquifer, while the other three (Agricultural, industrial and tourist sectors sustainably
developed; Energy used efficiently and in a sustainable manner; Measures to protected nature
formulated and applied) have an indirect relation to the karstic aquifers. A list comprising
25 different indicators was derived to assess pressures on groundwater quantity and quality
as well as resulting pressures on dependent ecosystems in selected transboundary aquifers.
Estimating indicator values will support sustainable water use and protection of ecosystems in
specific karst environment.
20.1 INTRODUCTION
Drivers Responses
Pressures
Impact
State
that links human activities to their ultimate environmental impacts and the societal
responses to these impacts.
The indicators in general should be used to measure achieved results of human
activities or to define the gap between desired and executed. For example, we may use
indicators to assess the status of environment and present geo- and bio diversity com-
ponents. An Indicator can also be defined as a parameter used to provide a measure
of the quality of water or condition of an ecosystem.
According to the UNECE report (2007) there are indicators of several types based
on the questions they answer:
Descriptive indicators answer the question: What are the pressures on the envi-
ronment and how is the quality of the environment developing?
Performance indicators answer the follow-up question: and is that relevant
for policy goals? Generally these indicators use the same variables as descriptive
indicators but are connected with target values.
Eco-efficiency indicators answer the question, Have we become more efficient in
our economic processes? Eco-efficiency indicators link driving forces with state
or pressure indicators. Efficiency refers to pressures such as emissions or natural
resources use.
Policy-effectiveness indicators answer the question, What has been the effect of
policy? (in terms of policy effects, for instance, reduced pollution due to strict
application of the principle Polluters pay).
Welfare indicators are connected with the question: and are we on the whole better
off? and ask for a balance between economic, social and environmental progress.
Smeets and Weterings (1999) also state that most countries and international
bodies currently develop performance indicators for monitoring their progress
towards environmental targets. These performance indicators may refer to different
kinds of reference conditions/values, such as:
Current Required
status status
Indicators
Visions or goals for the region: long-term targets that cannot be reached immedi-
ately but are needed to protect the environmental values of the water body. They
may be numerical, descriptive or relative statements (e.g. no net loss);
Resource condition targets: specific, measurable and time-bound targets used to
work towards the overall goal or vision for the water body. They may be medium
term targets (1020 years) required for regional planning (e.g. 30% reduction in
phosphorus loads by 2006); or
Management action targets: targets that are linked closely to management
actions. They are usually measurable and time-bound but with shorter time spans
(15years) than the regional targets (e.g. 70 km of riparian vegetation is to be
fenced within five years).
Main Target:
Environment
safe and clean
to be observed
Nature Human
activities
Quality and Quantity of: Settlements
(Large cities,
Touristic areas,
Water Islands,
(Aquifers, Villages)
Surface water bodies,
New sources...) Transport/Infrastructure
Air/Climate
Impact of: (Highways,
Road network,
(Air temperature, Airports,
Precipitation Railways,
Greenhouse gases...) Vehicles...)
Soil Industry
(Fertile land, (Small and med. enterprises,
Range land...) Manufacturers,
Factories)
Vegetation
(Grass, Agriculture
Forests, (Cropping,
Orchards...) Cattle breeding
Food production)
Fauna
(Endemic species, Energy
Protected species, (Mining,
Threatened species, TE Fossil fuels,
Wildlife, Hydropower,
Livestock, Nuclear,
Fishes...) Renewable energy)
Waste
(Solid waste,
Waste water
Hazardous waste)
Figure 20.3 Scheme of main nature components and human activities to be observed towards the
creation of a safe and clean environment.
292 Karst without Boundaries
environmental indicators are a key tool for environmental reporting and could be used as
a powerful tool to raise public awareness on environmental issues. Appropriately chosen
indicators, based on sufficient time-series data, can show key trends, help describe causes
and effects of environmental conditions, and track and evaluate policy implementation
(UNECE, 2003). Providing information on driving forces, impacts and policy responses
is a common strategy to strengthen public support for policy measures.
46.000000
46.000000
Hungary
Romania
Serbia
44.000000
44.000000
Ad
ri
at
ic
se
a
Legend
Boundaries
Kosovo*
State Boundary
42.000000
42.000000
Disputed State Boundary
Dinaric Karst
Diktas Boundary
F.Y.R.
Macedonia
Hydrogeological Units
Symbol Hydrogeological Description
Greece
40.000000
40.000000
Karst aquifers of moderate permeability, karst-fissure type of
KA2
porosity and medium level of transmissivity.
Figure 20.4 Distribution of main karstic aquifer systems in the Dinaric karst and DIKTAS project study area.
294 Karst without Boundaries
why the DIKTAS project was initiated and included three former Yugoslav countries
(Croatia, Bosnia & Herzegovina, Montenegro) and Albania.
After delineation of several transboundary aquifers (TBA) of major concern, fur-
ther analysis included the characterisation and development of conceptual models.
Groundwater budgeting of TBAs created a base for the assessment of groundwater
reserves and availability, as well as for proposals and measures aiming to ensure sus-
tainable development of TBAs. The analysis indicated that water extraction was still
far below the aquifers replenishment potential, and there is no evidence of significant
over-exploitation in the studied TBAs. For instance, in the case of Cetina and Neretva
TBAs, the average extraction of groundwater is ten times less than the total minimum
discharge of the springs (dynamic reserves). However, shortage of water is locally in
evidence during summer and early autumn months which coincides with increased
demands during the tourist season. That principles of the EU Water Framework
Directive (WFD) regarding ecological flow for downstream consumers has to be fully
respected, further complicates the water and environmental situation.
It is generally assumed that karst water quality is satisfactory even though
Dinaric karst aquifers are highly vulnerable to pollution. However, the catchments
are sparsely populated in the mountainous areas and there is virtually no intensive
farming or industrial activities. However, when pollutants are present (mines, indus-
trial and domestic waste waters, solid waste dumps, fertilisers), deterioration of water
quality in unconfined karstic aquifers is almost assured.
One of the tasks of the DIKTAS project was to prepare a proposal for the creation
of a new Groundwater Monitoring network which will fully respect karst specific
behaviour and include local water users (waterworks, dams, irrigation, industry). The
Cijevna/Cemi TBA shared between Montenegro and Albania has been identified as
the most problematic in terms of available data on water resources, and installation of
a modern monitoring network for observation of climate elements, and both surface
and groundwater has been proposed.
Therefore, the Dinaric region has a very dynamic water budget, one of most
intensive in all of Europe, but there are numerous challenges for sustainable utilisa-
tion of groundwater. These include high annual variation of natural flows and the
vulnerability of aquifers to pollution. It is thus important to (1) improve the quality
of water by eliminating or mitigating sources of pollution, (2) regulate the minimum
spring discharges, (3) ensure ecological flows, and (4) establish proper water moni-
toring systems. These actions are recommended as a priority during implementation
of the DIKTAS Strategic Action Plan (SAP).
To assist in attaining the vision for the Dinaric karst aquifer system, five overar-
ching Water Resources and Ecosystem Quality Objectives (WR/EQO) were suggested.
These are:
Groundwater Quantity
Water Resource/Ecosystem Objective A: To ensure sufficient groundwater availability
in dry periods, especially for water supply and to support environmental flow.
Groundwater Quality
Water Resource/Ecosystem Objective B: To maintain and improve (where required)
quality of karst groundwater in the Dinaric region.
Creating environmental impact indicators in dynamic karst system 295
A main objective of the DIKTAS project is to facilitate the equitable and sustain-
able utilisation of the transboundary water resources of the Dinaric Karst Aquifer
System, and protect the unique groundwater dependent ecosystems that character-
ise the Dinaric Karst region of the Balkan Peninsula, the environmental quality tar-
gets should be focused mainly on natural reserves, but all other environmental issues
which indirectly impact karstic aquifers, or depend on them, must also be taken into
consideration, evaluated and improved.
In total, five strategic Environment Quality Targets (EQT) are suggested. Two
have a direct, and three have an indirect relation to the karstic aquifers or are depen-
dent upon them:
Direct EQT:
1. Groundwater sustainably used.
2. Water and soil quality controlled and improved.
Indirect EQT:
3. Agricultural, industrial and touristic sectors sustainably developed.
4. Energy used efficiently and in sustainable manner.
5. Measures to protect nature formulated and applied.
To improve the environmental situation in the Dinaric karst region and achieve
the targets the following measures and activities are recommended:
Water must be used in a sustainable manner, by reducing consumption and losses
in current water utilities. Some new sources should be opened and some others reg-
ulated to improve their minimum flows and safeguard water dependent eco-systems.
The drinking water quality should be preserved from any kind of pollution, and
sanitary conditions improved in general: the air must be clean enough not to rep-
resent a risk to human health or to animals, plants or cultural assets. Solid waste
should be sustainably managed with greater emphasis on source reduction, recycling
and recovery. Waste water treatment should be widely applied in industry along
with the principle polluter pays. The energy saving and new sources of renewable
energy such as hydropower, solar, wind and sub-geothermal waters should become
dominant. Food production should be based mostly on organic agriculture with min-
imal use of pesticides and herbicides. Animal husbandry as an agricultural com-
ponent is essential to the long-term conservation and development of natural and
cultural assets in the agricultural landscape. The value of pasture land and forests
Creating environmental impact indicators in dynamic karst system 297
for biological production must be protected, at the same time as biological diversity
is safeguarded. Biological diversity must be preserved for the benefit of present and
future generations. Special attention should be paid to strict protection of threatened
and endemic species.
However, all these protection measures should not limit further development of
tourism and recreation but they have to be oriented and organised more ecologically
and environmentally.
Establishing a proper system for monitoring water quantity and quality, con-
tinually evaluating the status of environmental impact indicators, improving local
technical capacity and raising awareness of the local population on the importance
of water savings, protection from pollution, sustainable development and tourism,
remain major tasks beyond this phase of the DIKTAS project.
Not all of the EQT are easily and fully achievable because they largely depend on
national GDPs and economic growth, and they should have a long-term effect. For
this reason a phased approach is proposed, which, similar to the concept of the EU
Water Framework Directive, includes an interim stage of 15 years with partial EQTs.
Table 20.1 contains the list of these partial EQTs and related actions and indicators.
It is important to note that more attention has been paid to direct EQTs, because of
the main objective of DIKTAS.
UNECE (2003) stated that environmental monitoring systems are crucial for envi-
ronmental policy: they are the eyes and ears for policy makers, researchers, and the
public seeking to understand and improve the environment. National environmental
policies, institutions and funding mechanisms provide the context for monitoring sys-
tems. National policies can also specify goals for their development. Moreover, provid-
ing information to support national policies is a key objective of monitoring systems.
Water dependent ecosystems are essential components of the watersheds which
are under increasing pressure from human activities. In karst, dependent ecosystems
are exposed to greater potential hazard if they depend on water from aquifer. Although
the problem of aquifer over-exploitation is often exaggerated (Custodio, 1992; Burke
& Moench, 2000), variable water regime and low water flows during periods of
maximum demands (summer months) can cause stress in many aquatic systems. The
problem is much more sensitive when it comes to the area of transboundary concern
(Chilton, 2002; Puri & Aureli, 2005).
Table 20.1 Proposal of environmental quality targets and indicators.
Strategic EQO and Partial EQ targets to be achieved in next Main environmental impact
targets 15 years (up to 2030) Related actions indicators
Groundwater Created expert teams and completed WR WR surveying and assessing Renewable freshwater resources
sustainably used survey in designated areas in Dinaric karst and Promoting water savings, gradually increasing Renewable freshwater resources
TBA water taxes in recession (drought) periods
Reduced water spending by some 10% and Introducing into water practice aquifer engi- Domicile (and External)
specific consumption rate fell from ca 300 l/cap/ neering regulation to improve minimum flows freshwater resources
day to 270 l/cap/day (managing aquifer recharge) Renewable GW resources
Designed and implemented 6 pilot projects for Increasing efficiency and reducing losses in (Dynamic reserves)
aquifer engineering regulation and improved centralised waterworks; Renewable GW resources
minimum flows
Controlling fresh GW extraction especially (Dynamic reserves) in critical
Repaired and replaced pipelines in problematic along coastal strips and on islands; drought periods
waterworks and reduced losses in the region by
Increasing number of connections to water- Water exploitation index
5% on average
works and improving rural water supply. Groundwater exploitation index
Designed and implemented 4 pilot projects for
water recycling and reuse Water demands (availability)
Established proper GW monitoring network by Drinking water demands
systematic control of water extraction rate and Water available per capita
water quality by operators along with created Groundwater depletion
and equipped new GW monitoring sites
Water losses
Regional Consultation and Information
Exchange Body (CIE) created and put into
practice consultation / intern. reporting body
and mechanisms for equitable and sustainable
water use
Water and Major pollutants eliminated or emission of Controlling and inspecting systematically all Drinking water quality
soil quality pollution drastically reduced by putting into registered major pollution emitters Industry waste water index
controlled and practice the principle polluter pays Building new waste water treatment plants Household waste water index
improved Waste water treatment plants constructed in 30 Constructing environmentally safe and con- Specific pollutants index
cities and treatment of communal waste water trolled regional and local landfills Landfill status
increased by 30% in comparison with actual Cleaning and remediating illegal and improp- Water reuse
status erly built landfills and dump sites Salt water intrusion
Removing solid waste to the newly built and
Constructed 15 new regional landfills and 50
secure landfills
smaller landfills at municipality level. All landfills
Removing and preventing further depositing
are sanitary equipped and operated in a way
of all highly toxic substances at communal
that prevents adverse environmental impacts
landfills
Eliminated and cleaned sites of more than 300 Widely introducing household waste sep-
illegal small landfills and solid waste dump sites aration and recycling; Sending electronics,
with special attention to sanitation of karst electrical equipment and batteries for
depressions and features (small poljes, uvalas, recycling
dolines, potholes) Constructing new sewage systems in number
Increased number of connections to municipal of settlements of various size
sewage system by 40% Introducing and promoting advantages of
Reduced number of used septic tanks in water reuse and recycle
semi-urban and rural areas by 30% by con- Imposing system for regular water sanitary
structing new sewage systems control (constituents in accordance with
WFD, plus specific components putting water
Established functional system for water sanitary
at risk locally) and sharing information of all
control in all waterworks for centralised pota-
water utilities (National institutions, EIOnet
ble water supply which supplying more than
Water, DIKTAS area)
1000 users
Reduced by 30% fresh GW pumping in coastal
areas and islands during summer months to
prevent salt intrusions
(Continued)
Table 20.1 (Continued).
Strategic EQO and Partial EQ targets to be achieved in next Main environmental impact
targets 15 years (up to 2030) Related actions indicators
Created and put into practice new harmonised
legislation for sources sanitary protection and
zoning by emphasising specificity of karst
Established sanitary protection zones and
applied sanitary protection measures over all
sources for centralised potable water supply for
settlements over 1000 inhabitants
Constructing additional treatment segments for
potable water in ca 20 water utilities (filtration,
coagulation, removal of chemical contamination,
and if and where necessary, ozonisation versus
viruses, faecal coliforms)
Agricultural, Reduced by 50% use of pesticides, fertilisers and Producing organic food and limiting use of Fertiliser index
industrial and herbicides in agriculture agro-technical protective agents Pesticide index
touristic sectors Over-grazed pastures reduced by 10% Controlling and limiting use of compost and Irrigation water
sustainably sewage sludge that are applied to land for
Forested areas increased by about 10% demands and use
developed organic food production
Reduced use of nitrate based fertilisers in agri-
culture by 30% Constructing small waste water treatment
unit at all slaughterhouses or removing
Decreased concentration of nutrients in rivers,
organic wastes to nearby collecting /recycle
lakes and sea water by 30%
centers
Improved quality of littoral sea waters by 15%
Stimulating farming, keeping number of
Fish stock managed and harvested sustain- farms, farmers and livestock (large and small
ably and legally by applying ecosystem- based ruminants)
approach, so that overfishing is avoided
Conducting various erosion protective mea-
Economically used water from streams and sures, especially in areas with cultivated land
lakes and to a minimal extent from aquifer
Establishing new forest nurseries. Developing
systems for irrigation purposes. Groundwater
and implementing effective techniques for
pumping rate generally reduced by about 20% in
forest restoration
comparison with current stage
Imposedgreen practice concerning cleaning/ Systematically planting new trees
washing in all hotels in tourist resorts, reduced Widely introducing water saving technologies
specific consumption for ca 500 l/tourist/day to in irrigation such as drip or root irrigation
350 l/tourist/day
Fishing to be controlled by law and conducted
in sustainable manner
Building small biological treatment units at
resorts, hotels, restaurants and other touristic
infrastructure objects
Promoting green concept in main resorts
and informing tourists and visitors on
importance of water and energy saving and
supporting their environmentally friendly
behaviour during their stays and tours
Organising ecological and geo-heritage tours
focusing on karst features, wetlands, specific
biodiversity habitats and endemic species
Energy used Completed planned electro-energetic systems Completing environmental impact assess- Hydropower water use
efficiently and including dams and energy transmission lines ment studies for designed and newly planned Green energy contribution
in sustainable Installed appropriate filters at all air pollution energy objects and implementing all proposed
manner emitters (such as thermoelectric plants or big mitigation/monitoring actions
industrial factories) Constructing dams as well as water transfer
Controlled greenhouse gases emission and kept and energy transmission lines
at existing level Harmonizing and respecting legislation in
Introduced use of renewable vehicle fuels terms of maximaumgreenhouse gases emission
(biofuel) Controlling sulphur dioxide, nitrogen dioxide
Treated all mine waters before flowing to and oxides of nitrogen, particulate matter
nearest recipients (PM10 and PM 2,5), lead, benzene and carbon
monoxide in ambient air
Constructed numerous solar energy plants
and wind energy objects by increasing share of Constructing new hydro-energetic plants and
renewable energy by 10% in total energy sector numerous small hydropower plants on small
streams
(Continued)
Table 20.1 (Continued).
Strategic EQO and Partial EQ targets to be achieved in next Main environmental impact
targets 15 years (up to 2030) Related actions indicators
Increased number of installed heat pumps for Building solar and wind energy plants and
sub-geothermal water use by 100% compared implementing demonstration projects for sea
to actual status waves energy use
Constructed 2 pilot plants for conversion of sea Drilling and utilising groundwater for heat/
water waves into energy cooling purposes including dublet systems
(pumping/re-injecting water)
Stimulating use of solar panels for households
especially along coastal area and islands
Imposing stimulation rates for renewable
energy producers
Building local capacities and publicly promoting
green concept for energy saving and rational use
Measures to Reduced negative impacts of air pollution and Filtering gases emitted from industrial objects Protected habitat
protect nature noise disturbance of existing eco-systems by 30% and controlling noise in tourist resorts and in Water demands of dependent
formulated and Ensured ecological flows for all dependent off-roads in mountainous areas eco system
applied eco-systems by regulated minimal stream flows Managing aquifer recharge and regulating GW Specific endemic and endangered
and GW discharges at source sites discharge and stream flows species
Increased number of protected areas under Prohibiting hunting and controlling access to
various protective statuses and expanded pro- special reservation areas
tected land surface by 10%
Proposed selected karst features for protection Establishing criteria and forming teams and
as natural monuments and geo-heritage sites conducting surveys for selecting karstic fea-
due to their importance for humanity and karst tures of special importance to be protected
science (caves, potholes, uvalas, blind valleys, as geo-heritage sites, and proposing conserva-
karrenfelds, etc.) tion measures
Protected habitats and all threatened species Establishing system of control and protection
from IUCN red list of all endangered and endemic species in
Protected and monitored all endemic species karst areas and in karst underground
that inhabit the region, with special emphasis on
those in wetlands and caves
Creating environmental impact indicators in dynamic karst system 303
There are many references and projects related to environmental indicators which
cover different components of aquatic systems (including springs, streams, rivers,
lakes, wetlands, coastal lagoons and estuaries). Some of the more recent, such of Vrba
and Lipponen (2007) or UNECE (2007), pointed to a group of indicators helping to
evaluate pressures on water quantity and water quality.
Not all the indicators proposed (Table 20.2) have to be followed continually.
Those proposed for observation on an annual basis are as follows: Renewable
groundwater resources (item 1 from Table 20.2); Groundwater exploitation index
(5); Groundwater depletion (12). Others such as Specific pollutants index (17) and
Drinking water quality (by observing selected critical parameters), need more fre-
quent monitoring and the sampling frequency should be at least in accordance with
EU Water Framework Directive and European experiences (Jousma & Willems,
1996). However, many others should be observed continuously in an established
Groundwater Monitoring Network due to the specific intensive and variable
regime of Dinaric karstic aquifer systems, proportional to the complexity of the
status assessment of the groundwater body and presence of pollution trends. Most
of the monitoring sites should be located in drainage areas i.e. along base levels of
erosion and near recharge (ponors) and extraction sites (well fields, intakes).
Harmonisation of national legislation, legal and institutional reforms, creation of
a common or unique Water Information System and protocol for data exchange are
some of the proposals included as outputs in the Strategic Action Plan.
1 Water Renewable freshwater ratio: Total flow of surface and groundwater in the study area vs. mM3/year: mM3/year
Resources resources Total rainwater in study area (TBA catchment) or%
1a Availability Renewable freshwater Sub-indicator: As above but in critical drought periods mM3/4 critical months: mM3/4
(Pressures resources in recession (summer-autumn) critical months
on Water (drought) periods or%
2 Quantity) Domicile (and External)ratio: Total flow of surface and groundwater generated in the part mM3/year: mM3/year
freshwater resources of TBA inside each country vs. Total flow of surface and groundwa- or%
ter in the entire TBA catchment
3 Renewable GW resources ratio: Total flow of groundwater in the studied TBA catchment vs. mM3/year: mM3/year
(Dynamic reserves) Total rainwater in the studied TBA catchment or%
3a Renewable GW resources Sub-indicator: the same as above but in critical drought periods mM3/4 critical months: mM3/4
(Dynamic reserves) in (summer-autumn) critical months
critical periods or%
4 Water exploitation index ratio: Total water amount utilized for different purposes1 vs. Total mM3/year: mM3/year
renewable freshwater resources or%
5 Groundwater exploitation ratio: Total groundwater utilized for different purposes2 vs. Total mM3/year: mM3/year
index flow of groundwater in the study area or%
6 Water demands ratio: Total water demands for different purposes3 vs. Total renew- mM3/year: mM3/year
(availability) able freshwater resources or%
7 Drinking water demands ratio: Total water demands for drinking purpose vs. (1) Total renew- mM3/year: mM3/year
able freshwater resour-ces and vs. (2) Total flow of groundwater in or%
the study area
8 Water available per capita Water available (household water access) calculated per capita per m3/cap/
year year
9 Irrigation water demands ratio: Total water used for irrigation purpose vs. Total renewable mM3/year: mM3/year
and use freshwater resources or%
10 Hydropower water use ratio: Total water used for HP vs. Total renewable surface water mM3/year: mM3/year
resources or%
11 Green energy ratio: Summary green vs. total energy consumption and per every in millions Mw: millions Mw
contribution kind vs total or%,
12 Groundwater depletion Annual depletion of groundwater table (av. value) due to over m/year
abstraction. Punctually measured at selected points
13 Losses ratio: Total water losses (non-utilised)4 from the systems %
constructed for different purposes vs. Total tapped renewable
freshwater resources
14 Pressures Drinking water quality ratio: Number of samples of raw drinking water (from the sources) no: no
on Water with inappropriate quality5 vs. Total number of the controlled or%
Quality samples
15 Industry waste water ratio: Flow of untreated industrial (incl. mining) waste water mM3/year: mM3/year
index (returned to recipients) vs. Total flow of waste water generated in or%
study area
16 Household waste water ratio: Flow of untreated domestic waste water (returned to mM3/year: mM3/year
index recipients) vs. Total flow of domestic waste water in study area or%
17 Specific pollutants index ratio: Concentration (average) of selected component (pollutant) expressed in mg/l: mg/l (permitted
vs. maximal permitted level of the same component (pollutant)6 in level) or g/l: g/l (permitted level)
drinking water or % of samples of inappropriate
quality of cpec. comp. vs. total
samples
18 Fertiliser index ratio: Amount of mineral or organic fertiliers used per unit of kg/ha
arable land or tonnes /ha
19 Pesticide index ratio: Amount of pesticide used per unit of arable land kg/ha
20 Landfill status ratio: Number of inhabitants in study area without sanitary proper .000: .000
solid waste dumps vs. Total population in study area or%
21 Water reuse ratio: Reused or recycled water vs. Total flow of waste water in mM3/year: mM3/year
study area or%
22 Salt water intrusion ratio: Total water flow - already salty, brackish or under direct mM3/year: mM3/year
(in coastal aquifers) threat of intrusion vs. Total renewable freshwater resources or%
23 Protected habitat ratio: Total surface of protected area vs. Total surface of study area km2: km2
or%
24 Water demands of depen- ratio: Total water demands for downstream dependent eco system mM3/year: mM3/year
dent eco system vs. Total renewable freshwater resources-dynamic, or%
or Total water demands for (WDES) vs. Minimal discharge
(Continued)
Table 20.2 (Continued).
No Group Indicator Expressed as Unit
25 Specific species Specific water demands (flow) for endangered species throughout Presence of protected endemic
Sub-indicators: Specific the year (e.g. trout) species List
endemic and endangered
species (list)
1
Includes different end-users: Drinking water purpose; Irrigation; Industry; Hydropower; Water dependent eco-systems. The Indicator should be calculated for each
consumer separately, but also expressed as(1+2+3)vs. (5)
2
The same as above
3
Demands to be calculated for each specific end-user as in the case of items 4 and 5.
4
Note: Mostly referring to water transport. If water leaked from reservoir and is utilized downstream for another purpose this is not a loss.
5
No compliance with drinking water standards for whatever reasons (microbiology, chemistry, specific comp.)
6
Pollutant or specific component in concentration higher than permitted, such as NO3, P or PO3(4), pesticides, PCB, turbidity, biology indicators, etc. List to be specified in
accordance to actual situation within TBAs and in compliance with EU Water Frame Directive for surveillance and operational monitoring
Creating environmental impact indicators in dynamic karst system 307
The Guarani aquifer system is the largest groundwater resource in the world,
with 45,000 km3 of water and a surface area of 1.2 million km2 (Valente, 2002).
The transboundary aquifer is shared by Argentina, Brazil, Paraguay and Uruguay.
Approximately 24 million people live in the area delimited by the boundaries of the
aquifer and a total of 70 million people live in areas that directly or indirectly influ-
enced it. The main use of the aquifer is for drinking water supply, but there are also
industrial, agricultural irrigation and thermal tourism uses. In article 12 of their agree-
ment for sustainable utilization of GAS, the parties agreed to establish cooperation
programmes with the purpose of extending the technical and scientific knowledge on
the Guarani Aquifer, promoting the exchange of information and management prac-
tices, and developing joint projects, but no specific common monitoring programme
with identified critical parameters has been put in place.
The Genevese Aquiferextends over 19 kilometers underneath the southern
margin of Lake Geneva and the Rhne River across the border betweenFrance and
Switzerland. The width of the aquifer varies between 1 and 3.5 km. An average
of 1517 106m3of water is extracted annually from this aquifer. The two bilat-
eralagreements, in 1978 and 2007 (http://www.internationalwatersgovernance.com/
franco-swiss-genevese-aquifer.html), clearly define responsibilities, monitoring proce-
dures (extracted water quantity and water table variations, as well as water quality
control) and a reporting mechanism to the Management Commission for this aquifer.
(http://www.unece.org/env/water/meetings/legal_board/2010/annexes_groundwater_
paper/Arrangement_French_Swiss.pdf.). Although not karstic (the aquifer consists of
glacial and fluvio-glacial silt-sand and gravel), this TBA is an excellent example of
sustainable aquifer development and agreed proportional water share. By introducing
artificial recharge these two countries have also prevented aquifer over-exploitation
since the 1970s (Wohlwend, 2002).
Several more recent projects (e.g. GENESIS project, Preda et al., 2012) or
Commissions for large international watersheds (e.g. The International Sava River
Basin Commission 2011) classified the indicator packages or established a list of crit-
ical parameters recommended to be monitored.
Finally, DIKTAS Strategic Action Plan includes a proposal to establish a fully
equipped pilot monitoring site in Cemi/Cijevna karstic aquifer and also to further
expand the monitoring network in the other selected TBAs. The general setup for a
Monitoring Network should primarily include the following parameters (Stevanovic,
2014): Rainfall and other climate elements (air temperature, humidity, wind, evapo-
ration) observed on a daily basis. Riverflow observed on a daily basis limnigraphs
for automatic recording or gauging stations installed on major rivers and streams in
each country sharing TBA (entrance/exit stations). Springflow observed on a daily
basis as above, the limnigraphs for automatic recording or classic gauging stations
installed on major springs within TBA. Groundwater table observed on a daily basis
automatic data logger for groundwater table recording installed in piezometers prop-
erly selected to represent aquifer system in recharge/discharge areas in both countries
sharing TBA. In addition, a manual recording of the groundwater table on a daily/
weekly basis (depending on wet/dry seasons) should also take place on the piezome-
ters of the 2nd rank. Water quality control is to be organised in compliance with EU
Water Framework Directive requirements for surveillance and operational monitor-
ing. Sampling frequency and the number of observed parameters (salinity, chemistry,
308 Karst without Boundaries
turbidity, biology, specific components and pollutants) are to be adapted to local cir-
cumstances and pollution risks. As a minimum in the initial stage (surveillance) a set
of the complete analyses is to be organised on major springs, streams and piezometers
twice a year (high and low water periods).
To be able to define other environmental impact indicators in addition to the
above water parameters, relevant information on surface waters and groundwater
regime (quantity and quality) should be collected and provided on a regular basis to
the responsible authorities and local water management institutions such as water
agencies, hydrometeorological surveys, health and sanitary control centres, and
municipalities. Groundwater monitoring and data collection must be the task of all
those using groundwater for drinking purposes.
20.2 CONCLUSIONS
REFERENCES
Bonacci O. (1987) Karst Hydrology; with special reference to the Dinaric Karst. Springer-
Verlag, Berlin
Brundtland Commission (formally the World Commission on Environment and
Development,WCED), (1987) Our Common Future, Report. Oxford University Press.
Published as Annex to General Assembly document A/42/427, Development and International
Co-operation: Environment August 2, 1987.
Burke J.J., Moench H.M. (2000) Groundwater and society: Resources, tensions and opportuni-
ties. Spec ed. of DESA and ISET, UN public. ST/ESA/265, New York. 170 p.
Cvijic J. (1893) Das Karstphnomen. Versuch einer morphologischen Monographie.
Geographischen Abhandlung, Wien 5(3), 218329.
Chilton J. (2002) Preliminary assessment of transboundary groundwaters in South Eastern
Europe. UN/ECE Working Group on Monitoring and Assessment, Core Group on
Transboundary Groundwaters, INWEB. 15 p.
Creating environmental impact indicators in dynamic karst system 309
Preda E., Klve B., Kvrner J. et al., (2012) New indicators for assessing groundwater depen-
dent eco systems vulnerability. Delivereable 4.3. GENESIS FP 7 project: Groundwater and
dependent eco systems, pp.84, www.thegenesisproject.eu.
Puri S., Aureli A. (2005) Transboundary aquifers: A global program to assess, evaluate, and
develop policy. Ground Water 43(5), 661668.
Smeets E., Weterings R. (1999) Environmental indicators: Typology and overview. TNO
Centre for Strategy, Technology and Policy, The Netherlands, Eur. Environ. Agency,
Copenhagen, 19 p.
Stevanovic Z. (2011) Management of groundwater resources (in Serbian). University of
Belgrade, Belgrade, 340 p.
Stevanovic Z., Kukuric N., Treidel H., Peka Z., Jolovic B., Radojevic D., Pambuku A. (2012)
Characterization of TB aquifers in Dinaric karst A base for sustainable water management
at regional and local scale. Proceedings of 39 IAH Congress, Niagara Falls.
Stevanovic Z. (2014) Environmental impact indicators in systematic monitoring of karst aqui-
fer Dinaric karst case example, In: Kukuric N, Stevanovic Z, Kreic N (eds) Proceedings
of the DIKTAS Conference: Karst without boundaries, Trebinje, June 11-15 2014, 80-85.
Stritih J., Qirjo M., Cani E., Myftiu A., Spasojevic D., Stavric V., Markovic M., Simic D.,
Deda S., (2007) Environmental Policy in South-Eastern Europe. UNDP Report prepared for
Conference Environment for Europe, FSC. Belgrade. 240 p.
UNECE (UN Economic Commission for Europe) (2003) Environmental monitoring and report-
ing, Eastern Europe, the Caucasus and Central Asia, United Nations Publications, New York,
Geneva, 84 p.
UNECE (UN Economic Commission for Europe (2007a) Environmental indicators and indica-
tors-based assessment reports: Eastern Europe, Caucasus and Central Asia. United Nations
Publ. ECE / CEP 140, New York; Geneva. 93 p.
UNECE (UN Economic Commission for Europe) (2007b) Guidelines for developing national
strategies to use air and water quality monitoring as environmental policy tools for the
countries of Eastern Europe, the Caucasus, Central Asia and South-Eastern Europe, New
York; Geneva. p. 46.
Valente M. (2002). South America: MERCOSUR vows to take over huge water reserve,
Mercosur Article, Inter Press Service (IPS)/Global Information Network, July 22, 2002,
pp. 1-2.
Vrba J., Lipponen A. (2007) Groundwater resources sustainability indicators, IHP VI Series
on Groundwater No.14, UNESCO, Paris.
Water Framework Directive of EU, WFD 2000/60, Official Journal of EU, L 327/1, Brussels
Wohlwend J.B. (2002)An overview of groundwater in international law A case study: the
Franco-Swiss Genevese Aquifer, Workshop III on Harmonization of Diverging Interests in
the Use of Shared Water Resources, 17-19 Dec. 2002, Beirut, 1-24. Available at http://www.
bjwconsult.com/The%20Genevese%20Aquifer.pdf.
Chapter 21
ABSTRACT
Only 10% of the total groundwater resources in the southwest karst area of China is used.
Approaches to exploration, utilisation and management, should be based on good understand-
ing the karst aquifer. The palaeozoic carbonate rocks are characterised as old, hard and weakly
porous. Karst hydrogeological media are highly heterogenous with many different lithologies
including: limestone, dolomite, mixed limestone/dolomite. These lithologies are often interca-
lated or in alternation with clastic rocks. Spatial distribution of lithology impacts groundwater
flow. This chapter considers four case studies of karst groundwater exploitation by under-
ground dams: damming groundwater flow to form a groundwater reservoir and construction
of a grout curtain to stop the flow and transfer the groundwater to a reservoir.
21.1 INTRODUCTION
Carbonate strata are widespread across China and outcrop over an area of 3.44 Mkm2
(Datong and Yang, 1983; Datong 1985). The strata range in age from Precambrian to
Cenozoic. In southwest China there is over 0.53 Mkm2 of exposed carbonate terrain,
most concentrated in Guangxi, Guizhou and Yunnan provinces (Figure21.1) (Jianhua
and Daoxian, 2005). The distribution of the carbonate rocks in different provinces is
shown in Table 21.1.
The carbonate formations are mainly of Triassic, Permian, Carboniferous,
Devonian and Cambrian age. They occur over an area of 13.5, 10.2, 8.3, 7.9 and
6.6 104 km2 respectively, or 87.5% of the entire exposed carbonate in southwest
China (Table 21.1). The carbonate rock formations in the Cambrian, Cretaceous, and
Triassic supergroups mainly occur in Hubei, Guizhou and Sichuan provinces, while
the carbonates in the Carboniferous and Devonian mainly occur in Hunan, Guangxi
and Guangdong provinces (Figure21.2). The dolomite is, therefore, mainly distributed
in Hubei, Guizhou and Sichuan, representing 36.9%, 24.8% and 22.8% of the total
exposed carbonate rocks. Limestone is mainly distributed in Hunan, Guangxi, and
Guangdong, which occupy 68. 6%, 71.6% and 69.4% of the total exposed carbonate
rocks (Figure21.3). Therefore, the carbonate is characterised as old, hard and poorly
porous. The karst media is highly heterogeneous with different carbonate rock types
312 Karst without Boundaries
Legend
0 200 400 km
Figure 21.1 Distribution of karst terrains in southwest China (from Jianhua, 2005).
Table 21.1 The exposed area of carbonate rock (104 km2) from different geological era and in different
provinces in southwest China km2.
12
10
0
K-J T P C D O S Z
Geological age
Figure 21.2 Exposed carbonate rock area from different geological age in Southwest China.
Different carbonate rock proportion (%)
80.00 limestone
70.00 dolomite
60.00 limestonedolomite
50.00
40.00
30.00
20.00
10.00
0.00
ei
ou
an
an
an
xi
ng
in
ub
ng
hu
un
do
zh
gq
Yu
ua
H
Sic
H
ui
ng
on
G
G
ua
ch
Province
G
Figure 21.3 The proportion of carbonate rock types in various provinces in Southwest China.
typical high hydraulic conductivity of has a large impact on the entire hydrological
system. Groundwater flow in karst medium is significantly different from that in a
homogeneous porous medium, where groundwater flow velocity in conduits can be
much faster than that in homogeneous porous media. This will lead to faster ground-
water transport as well as contaminant dispersion (Daoxian, 1991; Andreo, et al.,
2010).
In southwest China a large part of the precipitation is rapidly transferred into the
underground system and flows down gradient after the rainfall event. Thus, despite high
precipitation and favourable water resources in this region, droughts are often reported,
particularly in the dry season. Furthermore, insufficient groundwater exploitation and
utilisation, as well as uneven spatial distribution of water, hinders regional economic
development (Riyuan et al., 2003). It has been reported by the Guangxi Department
of Water Resources that the total annual karst groundwater resource amounts to
314 Karst without Boundaries
48.4 billion m3, and groundwater extraction amounts to 1.274 billion m3, or only
2.6% of the total potential amount (Risheng & Shumin, 2006).
Low utilisation levels of karst groundwater in Guangxi province have constrained
domestic water supply, and water for industrial and agricultural development. Global
warming and extreme climate events increase the frequency of droughts and floods.
Exploitation and management of karst groundwater resources in southwest China is,
therefore, imperative.
Despite the difficulties in constructing underground artificial storage systems,
successful construction of underground dams in favorable hydrogeological and sub-
surface geological structures have been reported (Milanovic, 2004). Large numbers
of underground engineering structures were constructed in the southwest China karst
region to achieve better use the groundwater resources (Daoxian, 1991). Construction
of underground dams in the karst area requires massive and meticulous geological
and hydrogeological investigations, coupled with various geophysical techniques.
Four successful cases studies of karst groundwater exploitation by underground dams
in the southwest karst of China are presented.
15
3
D3y Cave
4
30
Ground dam
Underground
reserivor
Floodgate
Ground dam
Figure 21.4 Sketch of Fenfa underground reservoir connecting sections (from Mingzhang, 2006).
182.0
reservoir
sinkhole
limestone
drill
passage
concrete
Curtain grouting 1 N
(262 m)
Reservoir
18 00
Cu F3
2 25
5
rta
2
in sinkhole1 F2
gr
ou
tin F1
g 2 (144 m)
sinkhole4 sinkhole5
sinkhole2 (180 m)
80 sinkhole3 Curtain grouting 3 Cu
Legend rta
in g
rou
curtain tin (66 m)
g4
grouting
floodgate
falut
cave
Pool
fissure
altitute 70-8 0
0 32
sinkhole
drilling
hole
Pool
Figure 21.6 Location of sinkholes and curtain grouting in Dalongdong Reservoir (from Zhouxin, 1999).
A curtain grouting method was chosen to block water leakage. The grout curtain
extends to a depth at elevation 122.5 m. Four leakage grout curtain segments with total
lengths of 652 m were constructed from north to south of the hillside. The individual
length of the segments is 262 m, 144 m, 180 m and 66 m, respectively (Figure21.6).
Presently, Dalongdong Reservoir works well. It provides irrigating water for 1,200 ha
of farmland, and the associated hydropower production is 4.48 GWh (Figure21.7).
1506.7
623.0
516.5
P1 C3 C2 C1
dolomite of Permian and Triassic age. A highly developed conduit system occurs along
a fault zone in this area. Most of the water resources are underground water. The
drinking water and irrigating water of Suduku village mainly depends on seasonal
epikarst springs in the southeast of the village. In the dry season, potable water is insuf-
ficient, and sometimes there is a shortage of drinking water.
Hydrogeological investigations were carried out. The water flow path was in a
shallow zone of carbonate strata, and frequent crossflow of surface water to ground-
water was observed. The depth of the conduit bottom is about 12 m. The irregular
karst channel, is 1.8 m high and 2 m wide. The groundwater flow rate of 15.0 l/s.
A dam, 12 m high, was constructed using concrete to increase the ground
water level to intersect the ground surface for local water supply. The Suduku conduit
was blocked successfully to form an underground reservoir. The new supply greatly
improved the living standard of local residents. It was reported that the captured
groundwater flow rate is 1296 m3/s during the dry season. The volume of the dam is
30 m3 and it cost 50000 RMB (Figure21.8).
150
floodgate
farmland
underground
reservoir T2g
0 5 10 m
Figure 21.8 Sketch of the geological section of the Suduku underground dam (from Yu et al., 2006).
The elevation of the cultivated lands is much higher than that of the spring outlet.
Thus, the spring water could not be used to irrigate this cultivated land. However,
flood events happen frequently during the rainy season. In order to optimize ground-
water resources a grout curtain structure was used to increase the groundwater level
to irrigate the farmland.
Detailed karst hydrogeology investigations were conducted, with application of
remote sensing technology, high-density electrical method, groundwater monitoring,
hydrogeological boreholes and tracer tests. It was reported that the recharge area
of Pijiazhai Spring is 115 km2 in area, and 70100 m higher than that of Pijiazhai
Spring outlet. The hydraulic gradient is 1.431.62% and the water velocity is
144.82176.45 m/h. The Pijiazhai Spring water flow has hydro-powerful potential.
Therefore, the integrated dam structure was designed for water exploitation and util-
isation. It includes an underground anti-seepage grouting curtain, a Ushape dam to
form a small reservoir to concentrate flow, and pipeline system to connect the water
reservoir to a trench by adjusting water pressure and to release water for farmland
irrigation (Figure21.9).
After completion of the Pijiazhai system the groundwater level is now 4.4 m
higher than previously. The amount of groundwater which flows through the inverted
siphon to irrigate the farmland is 60000 m3/d. This structure resolved the agricultural
water shortage in the downstream farmland, and domestic water needs for potable
water in downstream area.
21.3 CONCLUSION
The spatial variation of carbonate rock in the different provinces in southwest China
creates varying karst topography. The four successful underground dam projects in
the karst areas provide illustrations of groundwater exploitation in southwest China:
damming methods including ground reservoir construction by blocking underground
Hydrogeological settings for underground dam construction 319
Water
Lead release
East channel sluice gate
Narrowing water
pressure control
pool
Anti-seepage
grouting curtain
1 2 3 4 5
Figure 21.9 Integrated dam engineering in Pijiazhai karst spring (from Yu, 2008). 1. karst condiut 2. soil
layer 3. water flow direction 4. previous water level 5. water level after damming.
conduit flow (in Guizhou); construction of grouting curtain to stop water leakage and
to transfer the karst flow into the surface reservoir (in Guangxi); damming the under-
ground flow located in the karst window to increase the water level to take water (in
Yunnan); and construction of U shaped concrete wall to block groundwater flow to
increase water level for irrigation of farmland (in Yunnan).
Successful retaining engineering structures in the karst area are very important for
local and regional socio-economic development.
REFERENCES
Andreo B., Carrasco F., Duran J.J., LaMoreaux J.W. (2010) Advances in Research in Karst
Media. Springer-Verlag Berlin Heidelberg.
Daoxian Y. (1991) Karst of China. China Geological Publishing House.
Datong L., Yang L. (1983) Measurement of Carbonate Rocks Distribution Area in China.
Carsologica Sinica 2(2), 147150 (in Chinese with English Abstract).
Datong L. (1985) Explanation of the map of soluble rock types in China. Cartographic
Publishing House (in Chinese).
Ford D., Williams P. (2007) Karst hydrogeology and geomorphology. Wiley, Chichester.
Goldscheider N. (2012) A holistic approach to groundwater protection and ecosystem services
in karst terrains. AQUA Mundi 2, 117124.
Jianhua C., Daoxian Y. (2005) Karst Ecosystem of Southwest China constrained by Geological
Setting. China Geological Publishing House. (in Chinese).
Milanovic P. (2004) Water Resources Engineering in Karst. CRC, Boca Raton.
Mingzhang W. (2006) Exploitation and Utilization of underground water resources in Guizhou
Karst Area. In: China Geological Survey, Institute of Karst Geology, CAGS (editors)
Exploitation and Utilization of Karst Groundwater Resources in Southwest China. China
Geological Publishing House. (in Chinese).
320 Karst without Boundaries
Risheng M., Shumin Q. (2006) Exploitation and Utilization of underground water resources in
Guangxi Karst Area. In: China Geological Survey, Institute of Karst Geology, CAGS (editors)
Exploitation and Utilization of Karst Groundwater Resources in Southwest China. China
Geological Publishing House (in Chinese).
Riyuan X., Jiansheng T., Yuanfeng Z. (2003) Sustainable utilization measures of groundwater
resources in karst areas of Southwest China. China Population, Resources and Environment
13(1), 8185 (in Chinese with English Abstract).
Yu W., Gui Z., Aihua L. (2006) Hydrogeological Setting for karst groundwater resources
exploitation and utilization in Yunan and some successful cases. In: China Geological Survey,
Institute of Karst Geology, CAGS (editors) Exploitation and Utilization of Karst Groundwater
Resources in Southwest China. China Geological Publishing House (in Chinese).
Yu W. (2008) Water resource exploitation from big karst spring by flow narrowing-water
pressure adjusting-water table raising A case in Pijiazhai spring, Luxi county, Yunnan,
Carsologica Sinica 17(1), 16 (in Chinese with English Abstract).
Zhuoxin W., Guangyan S. (1999) Treatments against the seepage of Dalongdong Karst
Reservoir in Shanglin County, Guangxi. Pearl River 4852 (in Chinese).
Chapter 22
ABSTRACT
One of the basic problems of dam and reservoir construction in karst is water tightness. As
a consequence of the nature of karst, and the associated insecurity of water storage due to
leakage, choosing optimal dam sites can be a highly complex task. Problems associated with
leakage are exemplified by the Viegrad dam site located in Bosnia & Herzegovina. During
the first year of the operation of this dam, the occurrence of submerged downstream springs
was noticed. Flow measurements of these springs established that the discharge increased from
~1.4 m3/s in 1990, to ~14.7 m3/s by 2013. This chapter presents results from field investigations
and modelling undertaken to establish the positions, geometry and flow through karst conduits
beneath the dam site. The modelling process consisted of formulating a 3D spatial model, and
subsequently a mathematical model of the groundwater flows observed in the system. Data
collected during subsequent remedial engineering works provided evidence that the method
applied for the construction of a 3D spatial model and subsequent parametric model of the
karst aquifer, aided by an incomplete data series, is feasible.
22.1 INTRODUCTION
conduits and pathways, and form a discharge outside the reservoir area. In response to
these findings, it is of paramount importance that, at an early stage of the site investi-
gation, the main emphasis must be on understanding the complex conditions of karst
groundwater circulation (Milanovic, 2015). Some inadequately explored dam sites and
reservoirs constructed in karst have never fully filled with water, for example the Hales
Bar dam in the USA or the Vrtac dam in Montenegro. Indeed, some have been com-
pletely abandoned after unsuccessful attempts to reduce water losses, such as at the
Montejaque dam in Spain. Some other dams have had sudden water losses even after
years of successful operation, or increases in leakage over time. Examples of this type
include Viegrad in Bosnia and Herzegovina; the Great Falls dam in Tennessee, USA;
the Mavrovo dam in FRY Macedonia; the Slano dam in Montenegro; the Hammam
Grouz dam in Algeria and others (Milanovic, 2000).
Romanov and Gabrovek (2003) state that karst conduits under natural condi-
tions are created on time scales of thousands to hundreds of thousands to millions
of years, as natural hydraulic heads drive water through narrow fractures which are
gradually widened by solution. By contrast, under reservoir condition where water
pressure of 5 to 10 or more Bar can accelerate conduit enlargement by more than ten
or one hundred times. For example, Bauer et al. (2005) demonstrated, through simu-
lation of a dual porosity system, that small conduits with initial diameters of 4 104
m can be enlarged by solution within tens of years and may cause serious water losses.
A study of karst conduit genesis and formation of a three dimensional spatial
model of the main karst pathways underlying the Viegrad dam site has been con-
ducted with the aim of better understanding and preventing leakage beneath the dam
site. The Viegrad hydropower plant is situated on the River Drina, 2.7 km upstream
from the town of Viegrad. It was built between 1985 and 1989. The dam of the
Viegrad Hydro-Power Plant is a concrete gravity dam. An integral part of the dam
is the 594 m long grouting curtain (325 m beneath the dam structure and 65 m in
the left abutment and 204 m in the right abutment) and 50 to 130 m deep. In order
to define the positions of karst conduits along which groundwater circulates under
the dam site, special-purpose investigations and remedial works were undertaken in
20092010 and again in 20132014. Remedial works are still in progress at the time
of writing (2015).
Geological investigations of the karst setting focused on a rather narrow area
containing a recharging sinking zone and a drainage discharge zone (Figure22.1).
The initial problem was how to perform a quality analysis to characterise the problem
sufficiently in order to enable effective remedial works to prevent leakage. The prob-
lem was approached theoretically, which initially played a major role and provided
guidelines for field activities; followed by detailed and complex field investigations; a
spatial 3D modeling of karst conduits; an empirical approach and later also a mathe-
matical approach aimed at producing the final form of the model.
The investigation area belongs to the Dinaricophiolitic zone, one of the most out-
standing and complex geological regions of the Dinaric karst. The dam site is located
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 323
Submerged springs
Reservoir
Visegrad
Dam
Figure22. 1 Visegrad dam; Photo above left submerged springs below the dam site during tracer test
(discharge zone); Photo above right reservoir (sinking zone); Photo below panoramic
view of Visegrad dam.
some are filled with calcite and terra-rossa. Due to high-pressure washing out of fine-
grained, incohesive sediments from the filled faults, the percolation process became
more intensive through time. The constant increase in the measured discharges
downstream from the dam, as well as the intermittent increased turbidity (from 5 to
65 NTU) of the water in the springs, indicates that the karstification process is intensi-
fied. According to inspection by divers, heavily clay-laden groundwater discharge was
emanating from the openings in the river bed downstream of the dam. Abrupt dis-
charge pulsation and pulsation of the quantities of mud in the discharging water was
also observed. The number of springs in the downstream river bed increased alongside
the discharge of water leaking beneath the dam. In the first five years of dam opera-
tion the leakage increased from approximately 1.4 m3/s to 6.5 m3/s, and subsequently
to 14.68 m3/s by 2014. The number of observed springs in the downstream bed of the
river increased from 3 to 18.
It is important to highlight that during the first filling of the reservoir, water (dis-
charging from the river-bed springs) was clear, but its temperature was higher than
average (14.5C18.7C), which indicates a process of forced circulation of hypogenic
water within the karst aquifer. Over time and with the increasing discharge capacity
of the karst system beneath the dam, the temperature of the downstream river bed
springs decreased to match the temperature of water at the bottom of the reservoir.
Geological investigations were carried out during all phases of design and construc-
tion of the dam for the Viegrad Hydro-Power Plant. The initial period of investiga-
tion had the primary goal of choosing the optimum dam site as well as definition of
possible seepages. These investigations have yielded a large body of results, a part of
which are relevant to the solution of the leakage problem under the present conditions
below the dam site.
Geological mapping of the catchment area of the Viegrad reservoir and dam site
was done at a scale of 1:10000, while an engineering geological map of the dam site
was produced at a scale of 1:500. More detailed geological mapping of the dam site
at scales of 1:50 to 1:1000 were also undertaken. More than 4470 m of drilling was
undertaken during the investigation, together with water pressure tests. Groundwater
tracer tests were carried out in the phase of choosing the optimal dam site. Systematic
monitoring has been undertaken from the stage of initial investigation and construc-
tion design to the present day, and leakage beneath the dam has been monitored from
1991 to the present day (Figure22.2) (IWD, Jaroslav Cerni, 2009).
In 1993 an emergency grouting programme was carried out to a depth of 110m
because of excessively large leakage flow. An initial attempt using polyurethane mixed
with cement and sand was unsuccessful. A second attempt involved injecting fast-set-
ting thixotropic cement containing 310 mm and 816 mm sand fractions and which
also included a proportion of fine cut plastic sponge. Generally this was unsuccessful
at reducing leakage, and leakage was still occurring through reactivated karstic con-
duits deeper than 110 m (Milanovic, 2004). During this period of remedial works
groundwater levels were monitored in the piezometers located in dam-abutments and
abutment injection galleries, and subsequently on a continuous basis throughout the
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 325
16
14
12
discharge m3/s
10
8
6
2
0
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
year
Figure22. 2 Diagram of water discharge (leakage) increasing through time (Milanovic, 2015).
1 2 3
52
Figure22. 3 Creation of a 3D tectonic model of the Visegrad dam area as one of the main input for
the construction of a karst conduit model network. 1. Fault detail with 3D position and
real inclination due to upper and lower layers 2. 3D model of two fault zone detail sketch
3. Fault network in investigation zone with non-permeable layer.
years 1993, 1994 and 1995. At present the groundwater level regime is monitored
twice a month in 58 piezometers.
A new period of investigation of the Viegrad dam began in 2009 specifically to
define the leakage flowpaths beneath the dam site. The first step was a detailed inves-
tigation of the narrower area around the dam site involving a geodetical survey of
the dam and the appurtenant structures on the scale of 1:1000 (IWD, Jaroslav Cerni,
2009). The second step was remote sensing investigations, based on the analysis of
satellite images and aerial photographic images from periods both before and after
construction of the reservoir. The main goal of these investigations was to determine
the fracture locations and morpho-structural texture of the rocks as the input data for
3D modeling of faults, including their dominant orientation, karstification and con-
trol on water circulation (Figure22.3). By combining the geodetic and remote sensing
data with existing results from geological mapping and coring, as well as the mapping
of the dam foundation and results obtained during the grouting program, more than
40 faults in the area of the dam were detected and classified.
The fault zones were assigned numerical designations to facilitate the definition of
their characteristics. The classification included the separation of regional and local as
well as less important structures. Structural blocks were identified based on the positions
and lengths of the structures, as well as fault widths and the positions and thicknesses
of lithological units mapped from boreholes. Larger structural blocks are restricted by
326 Karst without Boundaries
300
250
200
150
Legenda
Groundwater flow direction
Tracer injection zone and
time of appearance
GW in high karastification zone 100
Investigation borehole Karstic aquifer high karastified
Faults limestone
Low permeability rock
70 m n.m.
Figure22. 4 Hydrogeological cross section of Viegrad dam site with position of grout curtain and
investigation boreholes.
significant faults, and these large faults have widths such that cavities can readily be
enlarged due to flushing and chemical corrosion (i.e. karstified rock). Identified cavities
were classified based on derived tectonic characteristics (i.e., all faults and dominant
fractures within the dam area were classified based on their importance and size).
A second group of faults comprised of faults within larger structural blocks were
divided into two sub-groups major faults with wide zones, and structures with
smaller fault zones. The analysis was undertaken with satellite imagery analysis with
the main task being to identify the largest faults, primarily those located within the
area of the Viegrad Hydro Power Plant reservoir.
Geological investigations of the dam site started with new, detailed geological
mapping of the terrain (Figure22.4). Geological mapping of the site was the basis of
the investigation process, which included verification of previous data and additional
mapping. These materials were used to generate a detailed cross section of the dam
area (Figure22.4).
Cross-hole geoelectrical tomography scanning in the left and right dam abutment
and over the reservoir was carried out, as well as reflective seismic investigations
in dam galleries. The electrical tomography method was applied in two stages: first
between the boreholes BD-1 and BD-2 and then at the profiles between the boreholes
BD-1 and BD-4, BD-4 and BD-2 and BD-2 and BD-3. Following the completion of
both test stages (i.e., after geo-electrical tomography measurements at all four bore-
holes), a 2D model was created based on a finite element analysis, for the BD-3
BD-2 BD-4 BD-1 profile as a base for further 3D model forming.
Self-potential measurements were performed simultaneously in the reservoir
zone. Survey in the reservoir zone led to the detection of anomalous locations that
indicated possible water percolation. Two profiles were located in dam abutments
(left and right) and one located partly across the reservoir and partly across the left
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 327
dam abutment. The depth coverage was from 80 to 130140 m. Geo-electrical scan-
ning performed along dam abutments and the reservoir also indicated the positions
of anomalous geo-electrical environments with low electrical resistance values (under
60 Ohm m) and which were interpreted as fault structures through which water cir-
culation and water losses most likely occur.
At the locations of the anomalies, detailed diving investigations and underwater
video camera recordings were undertaken. This led to the discovery of a large sink-
hole at a depth of 50 m in the reservoir which was measured and at which the water
inlet velocities (0,5 m/s average) were determined (Figure22.5). Investigation of the
large sinkhole was made using underwater robots and other special hydrogeological
underwater equipment constructed by the Centre for Karst Hydrogeology, Faculty of
Mining and Geology, University of Belgrade. The sinkhole was also sed to investogate
spatial groundwater flows upstream from the dam by the misse a la masse method.
Downstream from the dam, in the part of the riverbed near the dam, a bathymetric
survey was performed, as well as further diving investigations and measurements of sub-
merged spring discharge. During this investigations divers explored 8 large and 12 small
springs with discharges from a few l/s up to a few m3/s (Qsum = 14.91 m3/s) (Figure22.5).
Exploratory drilling and corresponding investigations in the boreholes (down-
hole-video, geophysical logging etc.) were carried out. The locations of the boreholes
was determined from the results of all previous investigations. Video recording of
boreholes was one of the important contributors to the 3D spatial model. More than
350 individual karst features were recorded and correlated with previous geological,
tectonic and hydrogeological data (Figure22.6). The change in chemical composi-
tion of the water in the piezometers drilled at various positions in the reservoir and
downstream was more pronounced in the deepest levels 180230 m than in zone of
grout curtain 5080 m. This indicated that water flowing under the grout curtain was
Depth 52 m
SUBMERGED SPRINGS
LARGEST SPRING
Figure22. 5 Spatial position of sinkhole in the reservoir area and springs downstream of the dam and
some of underwater investigation results.
334 m n.m.
0 reservoir surface
zone of karstification conduit direction water
10 depth from surface karstification level and type azimuth circulation
GWL = 17 m 62 m horizontal fracture no
20 63 m low karstification no
30 68.5 m low karstification 145 low
68.9 m low karstification low
40 70 m low karstification low
50 botom of reservoir 74.2 m crack 120 low
7676.5 m kavern low
60
83.183.6 m low karstification no
Videoendoscopy after finishing 70 97.2104.5 m high karstified zone 198 low
borehole
105106 m high karstified zone no
80
106113.2 m high karstified zone no
90 120120.8 m low karstification no
124124.3 m high karstified zone 84 yes
100
Rotation camera for 124.3125.2 m kavern 7 yes
110 small diameter logging 125.2125.9 m kavern 194 yes
high karstified zone 126.7128.3 m kavern 222 yes
120
084 194 groundwater 135.3 m kavern 354 yes
130 354 185 flow direction
135.4136.3 low karstification no
007 222channel direction
138139 m kavern 47 yes
140
151.8156.2 m kavern 150/030 no
150 karstified zone
Entrance to the big kavern at 124 m 156.3158 m high karstified zone low
030 180 channel direction
160 160160.7 m high karstified zone low
161162 m karstified zone low
170 163.4 m karstified zone no
180 167 m karstified zone no
169.7 m slaba ispucalost no
190
172 m high karstified zone 182/045 yes
200 174.3175.1 m high karstified zone yes
186.6 m kavern low
210
karstifikovana zona 217 m high karstified zone no
220 040 234 220.4 m kavern 234/040 low
slow circulation 220.5223.6 m high karstified zone low
230 of groundwater
Karstified zone at 156 m 225.5228.2 m high karstified zone low
228.2232 m no karstification no
Figure22. 6 Photos of drilling rig position at the surface of the reservoir (top left); characteristic caverns identified in boreholes (middle and bottom
left); borehole log (centre) and summary data table of borehole RB 9 CCTV survey (right).
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 329
flowing faster than it was through the curtain itself. Similar results were established
during investigations of a leakage problem at Ataturk dam (Unal et al., 2007).
More than 40 tracer tests were carried out, employing in total more than 80 kg of
sodium fluoresceine and more than 500 kg of sodium chloride. Tracers were injected
in the existing piezometers upstream from the dam, then in known sinkholes, as well
as into the newly drilled boreholes. Importantly, tracer tests were repeated several
times in the large sinkhole and in some investigative boreholes as a basis for model
testing and calibrating.
The results of the tracer tests were generally indicative of possible directions
and lengths of karst conduits. The results of the tracer tests conducted in the large
sinkhole; boreholes BD-2, BD-3 and BD-4; and piezometers UD-42k and UD-45k,
demonstrated a complex network of rapid flow connections in the conduit system
beneath the grout curtain, connecting the large sinkhole with the springs in the River
Drina riverbed downstream of the dam (Figure22.7).
Results from the tracer tests in the large sinkhole at ~50 m depth showed that,
in every test, the first arrival of tracer was recorded after 29 minutes at the spring
located in the left portion of the riverbed (looking downstream, No. 17b) and then at
springs in the central portion of the riverbed after 31 to 34 minutes (in the following
order: springs 8, 7, 5, 3b and 4). Somewhat later, but with the longest duration of the
emergence wave, the tracer emerged at the largest spring (No. 1) after 36 minutes. The
final springs at which the tracer emerged (after 40 and 41 minutes) are located further
downstream, towards the right bank (13b and 18b).
Tracer tests conducted in the boreholes BD-4, BD-2 and BD-3 (located between
the right and central portions of the dam profile) and at piezometers UD-42k and
UD-45k, also confirmed the underground connection with all springs downstream
from the dam. In all tests with sodium chloride, as well as with sodium fluorescein,
the tracer discharged in the same order as in the case of the sinkhole tracer tests. The
main findings from the tracer tests are shown on Figure22.7 and Table22.1.
s
ing
pr
ds
ge
er
bm
su
re
ser
vo
ir
Sinkhole
17b 24 17 22 17 34 29 0.189
199 204 231 205 188 329
8 25 18 23 17 35 30 0.204
225 229 235 225 254 368
7 26 19 24 18 36 31 0.189
208 212 218 209 236 352
5 26 20 25 18 37 34 0.167
195 197 204 197 220 340
3b 26 20 25 20 38 34 0.161
178 182 190 186 205 328
4 26 20 25 20 38 34 0.170
195 220 208 203 223 346
1 30 22 28 22 41 36 0.191
262 265 275 270 288 412
18b 38 27 37 50 40 0.155
248 254 253 280 372
19 29 28 40 41 0.147
266 234 257 362
Tracer tests conducted from the right-hand bank of the reservoir resulted in low
tracer concentrations and long emergence times, suggesting a lower degree of karsti-
fication in this area.
It was considered appropriate to assume either that interactive work and inte-
grated use of known 2D and partly-defined 3D parameters was sufficient to produce
an output of a three-dimensional conduit defined within a 3D physical model. A con-
ceptual model based on the results of all of the investigations undertake is presented
schematically in Figure22.8.
Analysis of the geometry of the main karst conduits in the saturated zone below
the dam site and grout curtain connecting the reservoir with the discharge zone has
enabled the creation of a 3D model of karst channels. Analysis of the parameters
obtained through the quantitative and qualitative monitoring of groundwater charac-
teristics, and their analysis for the physical model, provided data on the relationship
between the recharge and discharge zones. Such a model was used for an analysis of
speleogenesis and the detailed hydrogeology of the dam area.
The basic problem in the determination of this methodology was how to
develop a model to further analyse the spatial position of karst conduits, integrating
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 331
HPP VISEGRAD
Reservoir
Right bank
Bottom water
Figure22. 8 Schematic longitudinal cross sections; 1 Location of first tracer appearance, 2 Location
of last tracer appearance, 3 Grout curtain, 4 Visegrad Dam, 5 Bottom Water,
6 Visegrad Reservoir, 7 Investigation borehole, 8 Karstified rock, 9 Limestone
with low karstification, 10 Groundwater flow direction, 11 Main ponor (sinkhole),
12 Channel with active water circulation, 13 Channel with slow water circulation due
to injection works.
Theoretical approach, which initially played a major role and provided guidelines
for field activities.
Detailed field investigations.
Development of a basic input 3D model, then an empirical approach and later
also a mathematical approach, aimed at producing the final form of the model.
The 3D geological model was developed for the purposes of generating a network
of potential karst conduits running from the identified infiltration zone to the accu-
rately defined karst discharge zone. The 3D geological model was developed using
ArcGIS software and the 3D Analyst, Spatial Analyst and Network Analyst extensions
(Milanovic, 2010). All spatial data, such as geological maps and profiles, as well as the
positions of the dam, grout curtain, grout galleries and piezometers, were converted
into digital format, and each spatial unit was defined by its x, y and z coordinates.
332 Karst without Boundaries
1 2
3 4
Figure22. 9 Steps of 3D spatial network generation; 1. 2D data layer with all collected data, 2. 3D mod-
eling of faults, 3. 3D modeling of all other data (boreholes, diving, tracer tests, geophysics
etc.), 4. Final formulation of 3D conduit network.
1. Individual fault digitisation, i.e. the creation of a series of 3D fault planes which,
in addition to the basic extent shown in the 2D presentation (plan view), are also
characterised by vertical attributes relating to depth and the angle of dip.
2. Generation of DEMs of the land surface and the base of karstification, as well as
3D models of faults whose outputs (nodes and lines) provide the starting point for
3D analysis.
3. Generation of an interactive node and line network based on faults, DEMs and
boreholes logs, as well as other directly quantifiable investigative data which were
integrated in a 3D spatial model of the karst conduit network.
The karst conduit network was generated by linking the most probable flow
directions, using tectonics as the safest parameter for the initial phase of evolution of
karst conduits over a certain time interval. A 3D network of potential karst conduit
pathways (Figure22.9) was constructed based on the results of GIS data processing.
The 3D model described the network of directions of potential karst conduits, from
the sinking zone to the discharge zone (Figure22.10).
As required by the numerical algorithms, this network was described by means of
model nodes and elements. They made up the topology of the numerical model and
spatially corresponded to the geological model. Nodes were represented via x, y and
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 333
RESERVOIR
DOWNSTREAM
RESERVOIR
Figure22. 10 Left 2D conduit network of possible leakage pathways below the dam site (model
result) Right 3D conduit network of possible leakage pathways below dam site
(Milanovic, 2015); 1 Dam, 2 Grout curtain, 3 Karst conduit, 4 Node at the sinking
zone, 5 Node at the intersection of fault (or on fault), 6 Node at the drainage zone,
7 General groundwater direction below dam site.
z coordinates, where potentials were computed. 1D elements were used for hydraulic
calculations based on the finite element method. The lengths of these elements were
defined by the corresponding nodes, and spatial mathematical calculations were per-
formed along them (Milanovic, 2015).
The network comprised 177 nodes linked by 226 elements (Figure22.10). Using
an optimisation algorithm, the number of nodes, elements and free parameters is
indicative of the complexity of the problem. As a result, 1130 parameters were deter-
mined and at the same time more than 100 computed values were compared with
measured values. This optimisation required substantial processor time and a parallel
genetic algorithm was, therefore, used for the final calculation.
The principal objective of the mathematical model was to determine the spatial layout
of the principal karst conduits and their physical characteristics. Taking into account
the satisfactory degree of congruence with the real system, the model defines all the
main groundwater flows, with the corresponding parameters: dimensions, resistances,
potentials, velocities and discharges. This model, with its results, represents a founda-
tion for the interpretation of geological data and the determination of the direction of
development of the dominant karst channels, i.e. water circulation paths. The mathe-
matical model was based on known physical dependencies which approximately define
the behaviour of the real system. Some model parameters, which affect model perfor-
mance, were unknown and their values were assumed from a set of possible values.
By simulating the adopted topology and flow characteristics in the hydraulic
model using all of the hydraulic parameters that had been obtained, an analysis and
definition of the dominant directions of water flow beneath the dam body can be
undertaken (Figure22.11).
The results of these investigations show that the greatest amount of water flows
from the reservoir underground through the large sinkhole which is best modelled
334 Karst without Boundaries
m3/s m2
10 50
25
5
0
0
Figure22. 11 Left calculation values of discharges along modeled karst conduits (m3/s). Right
calculation dimensions of underground flows represented as the areas of a transverse
section in m2 (IWD, Jaroslav C erni, 2009).
with length ~60 m and average cross-sectional area of 7.1 m2. According to the
measurement results ~8.15 m3/s infiltrates into the groundwater system through
this feature, with another ~6.5 m3/s infiltrating elsewhere in the reservoir bed.
From the large sinkhole are formed two karst conduits through which groundwater
flows.
Conduits on the left side, through which the water flows downwards from an
elevation of 260 m asl to an elevation of 235 m asl, have an average transversal
section area of 19.8 m2. Through these water flows with an average velocity of
~0.2 m/s, and the total discharge through this channel is ~4.1 m3/s. Along conduits
with a spatial position in the middle of the conduit net direction, flows underground
are 5.5 m3/s. The area of the transverse sections of these channels amounts to on
average 23.9 m2. The erosion process along the channel of diameter 14.4 m2, up to an
elevation of 228 m asl, the underground flow bifurcates into two arms. The average
area of the transverse section near the borehole profiles amounts to ~10 m2, with an
average water flow velocity of ~0.3 m/s. At the elevation of 147 m asl this profile
bifurcates, and the major portion of the water flows towards the middle part of the
riverbed toward springs 17, 8 and 1, where, at an elevation of 130 m asl, it joins
the smaller channel to the right. Flows ~2.5 m3/s occur in the channel that runs
from the left abutment, when the total discharge amounts to 3.8 m3/s, established
through 3D modelling and mathematical modelling of the karst conduit system
(IWD, Jaroslav Cerni, 2009).
22.6 CONCLUSION
The investigations defined the underground water courses and parameters necessary
for development of the technical solution for the permanent rehabilitation of seepage.
However, the sealing rehabilitation works under such complex conditions are com-
plex, and must be followed by detailed monitoring.
3D Conduit modelling of leakage below a dam situated in highly karstified rocks 335
The results of investigations show that the largest amount of water flows from the
reservoir underground through the bottom and through the main sinkhole with area
of 7.1 m2. The left direction of underground flow, i.e. the karst channel, was formed
along the dominant longitudinal fault structures. Along this direction is a system of
karst channels with two courses, that generally run towards the boundary of the dam
blocks 8 and 9. The right direction of the underground flow is along the fault struc-
ture and water flows through it generally towards block 5 of the Viegrad Dam. This
direction upstream from the dam bifurcates into two courses; one of them, the more
important one, passes under block 5, while the other turns towards the central part
of the riverbed. These two courses are joined downstream from the dam axis. In the
region downstream from the dam, underground courses are mainly developed along
the directions of the longitudinal faults closer to the left abutment. The connection
with the springs in the River Drina riverbed is along the axis of the diagonal faults.
The 3D and mathematical modelling of the spatial position of conduits and flows
will serve as a foundation for the simulation of technological processes in the process
of rehabilitation regarding water seepage.
The 3D position of karst conduits and parameters necessary for the development
of technical solutions for the permanent rehabilitation of seepage, were defined by the
investigations. However, the sealing rehabilitation works are complex, and detailed
post-remediation monitoring is required. This is necessary in order to make correct
decisions during and after remediation.
The new data collected during grouting works (20122014) showed that the
method which was applied for the construction of a 3D spatial (physical) model and
the model of the karst aquifer, aided by an incomplete data series, was useful. The
underground cavity is now filled with more than 37000 m3 of inert material (sand
fractions from 4 to 32 mm) and leakage has been reduced by up to 65%.
REFERENCES
Bruce D.A. (2003) Sealing of massive water inflows through karst by grouting: principles and
practice. In: Back B (ed), Sinkholes and the Engineering and Environmental Impacts of Karst.
ASCE Geotech. Spec. publ. No.122.
Bauer S., Birk S., Liedl R., Sauter M. (2005) Simulation of karst aquifer genesis using a double
permeable approach Investigations for confined and unconfined settings, Processes of
Speleogenesis: A modeling approach, Editor Gabrovsek F., ZRC SAZU, Postojna.
Bonacci O., Roje-Bonacci T. (2008) Water losses from the Ricice reservoir built in the Dinaric
karst, Engineering Geology 99 (2008) Elsevier, pp 121127.
Borghi A., Renard P., Jenni S. (2012) A pseudo-genetic stochastic model to generate karstic
networks. Journal of Hydrology 414415, 516529.
Fazeli M. A. (2007) Construction of grout curtain in karstic environment case study: Salman
Farsi dam. Environmental Geology 51.
Filipponi M. (2009) Spatial Analysis of Karst Conduit Networks and Determination of
Parameters Controlling the Speleogenesis along Preferential Lithostratigraphic, Horizons,
thse no 4376, Suisse.
Ford D., Williams P. (2007) Karst hydrogeology and geomorphology. Wiley.
Institute for Developing of Water Resources (IWD) Jaroslav Cerni, 2009., Design on rehabili-
tation regarding water seepage beneath the dam of the Viegrad hydropower plant, Summary
report on performed investigations, Belgrade 2009.
336 Karst without Boundaries
Milanovic P. (2000) Geological Engineering in Karst. Monograph, Zebra Publ. Ltd, Belgrade.
Milanovic, P. (2004) Water resources engineering in karst, CRC Press, 143 p.
Milanovic S. (2007) Results of recording of karstic features Ourkiss dam project. Report.
Hidrotehnika, Belgrade.
Milanovic S. (2009) Report on special investigation on dam site Viegrad, Inst. for Develop. of
Water Resources Jaroslav Cerni, Belgrade.
Milanovic S. (2010) Creation of physical model of karstic aquifer on example of Beljanica mt.
(eastern Serbia), Doc. dissert, FMG, University of Belgrade, Beograd.
Milanovic S. (2015) Choosing optimal dam sites and preventing leakage from reservoirs, In:
Karst Aquifers - Characterisation and Engineering. (Ed. Z. Stevanovic), Springer, Professional
Practice in Earth Sciences, pp. 531549.
Romanov D., Gabrovek F., Dreybrot W. (2003) Dam sites in soluble rocks: a model of increas-
ing leakage by dissolutional widening of fractures beneath a dam. Engineering Geology 70,
1735.
Romanov D., Kaufmann G., Hiller T. (2010) Karstification of aquifers interspersed with
non-soluble rocks: From basic principles towards case studies, Engineering Geology 116,
261273, Elsevier.
Stevanovic Z., Milanovic S., Ristic V. (2010) Supportive methods for assessing effective poros-
ity and regulating karst aquifers. Acta Carsologica, 39(2), 313329.
Therond R. (1972) Recherche sur letancheite des lacs de barrage en pays karstique. Eyrolles,
Paris.
Turkmen S. (2003) Treatment of the seepage problems at the Kalecik Dam (Turkey). Engineering
Geology 68, 159169.
Unal B., ErenM., Yalcin G. (2007) Investigation of leakage at Ataturk dam and hydroelectric
power plant by means of hydrometric measurements, Engineering Geology 93, 4563.
Zogovic D. (1980) Some methodological aspects of hydrogeological analysis related to dam and
reservoir construction in karst. Proceedings of 6th Yugoslav Symposium for Hydrogeology
and Engineering Geology, Portoroz.
Chapter 23
ABSTRACT
Mine dewatering operations in the Transdanubian Mountains, Hungary between 1960 and
1990 caused significant groundwater depressurisation and drying up of several karst springs in
the city of Tata. Following the termination of mining operations, the groundwater flow system
started to recover. Many of the former springs reactivated and further springs are expected
to reappear in the future, causing environmental issues. Spring locations in Tata are aligned
with deep tectonic structures both in uncovered and confined karst areas. The analysis of well
hydrographs indicates that there is no hydraulic connection between shallow and karst ground-
water bodies. The prediction of karst water levels based on physical curve fitting suggests
that equilibrium karst water level will be reached around the year 2018 at approximately
140 m asl. The chemical composition of karst waters in the Tata region confirms that they come
from a dolomitic aquifer. While the karst waters of the region show a uniform composition,
shallow groundwater shows a variable composition. The chemical composition of most reac-
tivating springs indicates karst water origin. Some springs discharge shallow groundwater and
show signs of local pollution.
23.1 INTRODUCTION
The area of interest is located on the northern edge of the Transdanubian Mountains,
Hungary (Figure23.1). The Tata Springs represent the natural outlet of the Transdanubian
karst aquifer. The aquifer is situated in Triassic limestones and dolomites. The
Transdanubian karst system was strongly affected by mine dewatering related to baux-
ite and coal mining from the beginning of the 1950s (Figure23.2). The total groundwa-
ter abstraction rate reached 12 m3/s during the period between the late 1960s and the
late 1980s. The intense karst water abstraction caused regional groundwater recession
(VITUKI, 2000; Csepregi, 2007), and as a consequence, several springs in the Tata area
disappeared during this period. Following the termination of mining operations in the
early 1990s, mining related groundwater abstraction significantly decreased, and total
abstraction rates had dropped to 3.5 m3/s by the late 2000s . As a consequence the flow
system started to recover. Since the late 1990s the karst water table has risen by more
than 40 m in the Tata area. As a result, some of the former springs reactivated and fur-
ther springs are expected to reappear (Maller & Hajnal, 2013).
During the 1970s significant developments took place in areas that were pre-
viously used as agricultural land. Currently 30% of the population of Tata lives in
338 Karst without Boundaries
270000
TATA
250000
230000
210000
Figure23.1 Site location with dominant groundwater flow directions in natural state.
this area. The reactivating springs cause significant environmental problems related
to sewerage and water quality.
The aim of the study was to understand the hydraulic and hydrogeochemical
behaviour of the recovering flow system, to delineate affected areas and to provide
predictions on the location and timing of spring reactivation.
The karst springs at Tata represent one of the main natural outlets of the Transdanubian
carbonate aquifer system. The regional erosion base is the ltal-r creek. The recharge
areas of the springs are located in the north western uncovered carbonate aquifers of
the Gerecse and Vrtes Mountains, and in the north eastern karst areas of the Bakony
Mountains (Csepregi, 2002) (Figure23.1).
Reactivation of karst springs after regional mine dewatering 339
15
Groundwater abstraction (m3/s)
10
0
1951 1954 1957 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 2014
year
Budapest m3/s DKH K-i rsz Tatabnya Dorog
Kincses Vrpalota DKH Ny-i rsz Nyird, Halimba
Ajka Balinka, Dudar Forrsok Hvzi-t
Figure23.2 Total groundwater extracted by minng operations in the Transdanubian Mountains after
(Csepregi, 2007),
The karst springs in the Tata area are located along a chessboard-like fault system
which developed during tectonic movements between the Eocene and Pliocene periods
(Figure23.3). Before the beginning of mine dewatering operations, there were several
active springs in the area, at topographic elevations between 118141 m asl with
yields between 0.00001 and 1 m3/s (Horusitzky, 1923). Spring locations are aligned
with deep tectonic structures both in unconfined and confined karst areas. Although
there is little evidence of karstification, concentrated groundwater flow takes place
along tectonic structures representing preferential flowpaths. The confining marls in
the city of Tata do not block groundwater discharge to the surface along tectonic
features.
Potentiometric data in the monitoring wells indicate a very flat karst water sur-
face. The hydraulic gradient is in the range of 0.0010.0005. The natural regional
flow direction in the carbonate basement was from southwest to northeast. As a con-
sequence of mine dewatering, the natural flow directions were altered because of a
large depression located south east of the Tata area.
The karst water level in the vicinity of Tata was around 136 m asl (in the obser-
vation wells T-2: 135.6 m, T-3: 135.5 m, Lo Presti: 136.0 m, Tkr: 136.2 m,
Fnyes-1: 138.3 m; locations shown on Figure23.3.) based on data from January
2014. According to the low hydraulic gradients throughout the investigation area
the karst water level can be approximated with a uniform value of 136 m asl at the
beginning of 2014.
The original, undisturbed karstic water level data differs between various refer-
ences, and vary between 140 and 143.5 m asl (Horusitzky, 1923; Fogarasi, 2001). A
modelled water level was calculated at 138 m asl for 2020 by Csepregi (2002). The
natural-state water level data involves uncertainties. It is assumed that the karstic
level was about 140 m asl near Tata based on historical spring levels, and expect a
similar water level by the end of water level recovery, knowing that this also depends
on water extraction and on future climate conditions.
595000
Legend
Observation bores
Springs
tectonic zones
ltal-r Creek
Tata Lake
soPa2
sPa1-2-soPa2
kPa1-2
sPa1-2
c-mOl1-2
255000
255000
cOl1-2
taK2
J
dT3
595000
Figure23.3 Geological settings and locations of springs and monitoring bores. Pa1-2 and Ol1-2 indi-
cate confining Pannonian (Upper Miocene Pliocene) and Oligocene sediments. T3J and
K2 indicate the unconfined zones of the Upper Triassic, Jurassic and Upper Cretaceous
carbonate aquifers.
Reactivation of karst springs after regional mine dewatering 341
A groundwater level monitoring system has been operated by the regional Water
Directorate for several decades, and this has been expanded with new auxiliary karstic
and shallow groundwater level monitoring wells and contamination monitoring wells
during the implementation of the safety regulations for drinking water protection
areas in 2002 (Csepregi, 2002).
137
Tkr
136
Lo Presti
135
Tszf-10
134
133
T-2 T-3
132
Tszf-2/a
131
130
measured water level (m ASL)
129
Tszf-1/a
128
127
126 Tszf-2/b
Tszf-2/b
125
Tszf-3/a
124
123 Tszf-3/a
122 Tszf-2/a Tszf-9
121
Pokol
120
T-1
119 Tszf-7
Tszf-8
118
Fnyes
117
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
T-1 T-2 T-3 Tszf1/a Tszf-2/a Tszf-2/b Tszf-3/a Tszf-7
Tszf-8 Tszf-9 Tszf-10 Fnyes Pokol Lo Presti Tkr
Figure23.4 Time series water level data for monitoring wells at Tata.
342 Karst without Boundaries
The changes in groundwater levels in the monitoring wells between 2001 and
2014 are shown in Figure23.4. Wells T-1, T-2 and T-3 were sited to monitor the
changes in groundwater level and groundwater chemistry in the main karst aquifer.
Wells Tszf-1 to Tszf-18 were installed to characterise the changes in groundwater level
and groundwater chemistry of shallow groundwater, and to characterise the hydraulic
connection between shallow groundwater and karst water.
One of the practical goals of the study was to predict when the flow system reaches
natural equilibrium. The karst water level prediction was based on time series of
several karst water monitoring wells. A logarithmic trend line was fitted on the time
series of wells T-1, T-2 and T-3. The application of a logarithmic function for
curve fitting was based on the assumption that the recovery follows the Cooper and
Jacob (1946) well function. The Cooper and Jacob solution is an approximation of
the Theis (1935) non-equilibrium method.
Curve fitting was performed for the 20012009 period (Figure23.5), since the
extremely high precipitation in 2010 broke the trend of previous years, causing more
than 4 m rise in karst water levels. The trend-line was shifted by 4 m in 2010 to repre-
sent the trend characteristic of the following years. The fitted equation and the karst
water prognosis up to 2020 are shown in Figure23.5. The prediction based on curve
fitting suggests that the equilibrium level of the karst water table in the study area will
be reached around 2018 at approximately 140 m asl. This prediction is based on the
assumption that water abstraction rates and climatic conditions recorded between
2000 and 2009 remain constant during the following years. Any significant change
in these parameters might influence the recovery process and thus the time of spring
reactivation.
The horizontal dashed lines in Figure23.5 show the topographic levels of the
springs. The intersection of these topographic levels with the fitted trend line deter-
mines the predicted date of reactivation of the surveyed springs, and also provides
information about expected karst water level changes until 2020.
While most springs have already reactivated during the past years, the Lo Presti,
Kismos, Pokol and Tkr springs are expected to reactivate until 2018. The water
level in the wells located at higher topographic levels (Piarista well, Barta well, Piactri
well) are predicted to rise by a further 45 m by 2018.
After the end of karst water recovery any oscillations in water levels will be
determined by climatic conditions. These water level variations are expected to range
between 0 and 2 m.
Locations of natural springs and dug wells are indicated in Figure23.6. The 136 m
asl topographic level corresponds to the current karst water level, while the predicted
equilibrium karst water level is around 140 m asl. The grey area in between the two
isolines is subject to the possible reappearance of the springs in later years. The striped
areas indicate tectonic zones assumed to represent preferential pathways for ground-
water flow. The outcrop line of the Mesozoic carbonate aquifer is indicated with a
grey dashed line.
145
Piarista Kt (143,5)
Pokol F. (140,8 m)
140
Lo Presti F. (139,3 m)
Kis mos Kt (139,3 m)
y = 120,37In(x)-1144,5
Tkr F. (138 m)
135
Karstwater Ievel (m)
Trkfrd F. (135,3 m)
Zsid Iskola Ktja (133 m)
Piactri Kt (131 m)
130
Minnich Kt (130,3 m)
Vadszbolt (127,8)
Mjus 1 t 43 (127 m)
125
Lelkes F. (125 m)
T_1
120
T_2
T_3
trend 20112014
115
2000 2003 2006 2009 2012 2015 2018 trend 20012009
Time (year)
Figure23.5 Karst water level prognosis based on time series of the karst water monitoring wells in the Tata region and the topographic levels of
main springs.
595000
Legend N
Observation bores
Extraction bores
Springs
Aquifer outcrop boundary
tectonic zones
2014-2018 groundwater recovery zone
ltal-r Creek
Tata Lake
255000
255000
595000
Figure23.6 Locations of natural springs and dug wells together with topographic elevation contours.
Shaded area indicates the future zone of groundwater heads rising above ground surface.
Reactivation of karst springs after regional mine dewatering 345
TATA 26 Trkfrd
TATA 27/A Kastlykert
80 80 Bds csorg
TATA 28_m1
>
<=
40 40 TATA 41 Vadszbolt
+M
60 60
40 40
20 20
80
60
60
80
40
20
40
20
Ca Na+K HCO3+CO3 CI
Figure23.7 Chemical characteristics of groundwater based on data of karst water wells, springs, see
pages, a stream and a dug well.
346 Karst without Boundaries
1983) based on the changes in the chemical composition due to the effects of mining
activities. Medians were calculated for both data groups.
The chemical composition of the surveyed wells, the Fnyes, Lo Presti, Trkfrdo,
Pokol springs and the new Attila seepage and seepages at Mjus 1.u 43 and 45 show
karst water composition with CaMg-HCO3 or MgCa-HCO3 water types.
The chemical composition of Kastlykert spring and of the dug well at Kismos
stream are similar, both having a CaMg-HCO3SO4 water type. This shows the effect
of mixing with shallow groundwater and potentially also of local pollution. The high
nitrate concentration in the Kastlykert spring also suggests pollution from an anthro-
pogenic source.
Sampling sites Bds csorg and Lelkes seepages have a distinct chemical compo-
sition of the MgCa-SO4HCO3-MgCaNa-SO4HCO3 water type.
Kismos stream as a local discharge area, does not show any connection with the
karst water, is of the CaMg-SO4 water type, and is badly polluted.
Figure23.8 illustrates the data obtained from chemical analyses for groundwater
from shallow bores. The chemistry of shallow groundwater (TSZF wells) is typically
of CaMg-HCO3SO4 type, but some of them are of NaCaMg-HCO3 or CaMg-SO4
type. Their water composition covers a much larger range than the karst waters,
reflecting the effect of the local near surface geology and hydrogeology, which is
probably influenced by local tectonics. While the karst waters of the region show a
uniform composition, the shallow groundwater shows a variable composition. Many
of the shallow groundwater wells, not just the local dug wells, are polluted with
nitrate. This pollution is local, and usually cannot be found in the adjacent wells or
springs suggesting that nitrate might originate from septic tanks or a leaking sewage
system.
TSZF-1/a
80 80 TSZF-2/a
=>
O4
TSZF-2/b
<=
60 60
Ca
+S
TSZF-3/a
40
+M
40
CI
TSZF-4
g
20 20 TSZF-5
TSZF-6
Mg SO4 TSZF-7
TSZF-8
80 80 TSZF-9
60 60 TSZF-10
TSZF-14
40 40
TSZF-17
20 20 TSZF-18
80
20
40
60
80
20
40
60
Ca Na+K HCO3+CO3 CI
The changes in chemical composition through time are shown for some parameters in
Figures 23.923.11. The hydrogen-bicarbonate content of karstic waters is typically
400 to 500 mg/l (Figure23.9), while in the shallow groundwater (TSZF wells) varies
in the range of 100 to 800 mg/l (not shown in figure). The effect of mining activities
on the karst water composition is shown in Figure23.9 where the hydrogen-bicar-
bonate data of some representative karst wells and springs are plotted against time.
It can be seen that the hydrogen-bicarbonate content dropped at the beginning of
1983, with median values decreasing from 476 mg/l to 458 mg/l. This might indicate
a hydrochemical response to aquifer dewatering which started in the early 1950s and
intensified around 1972 to 1973 in the Tatabnya region.
The recent data from new karst water seepages (43 Mjus 1 street. 45, Mjus 1
street) also show these decreased concentrations, but samples collected from Lo Presti
karst spring and the two drinking water supply wells (Tata 28, Tata 34) in the last quar-
ter of 2014 and in 2015 show slightly increasing concentrations (median 3 = 468 mg/l).
These minor concentration changes are within measurement error limits, but the general
pattern, including concentration changes in other anions, suggests a recent change in karst
water composition starting probably from the mid-2000s, and increased since 2010.
550 140
1.
Tata 28_Fnyes Vzmu Mjus 1.u. 43.
2.
Tata 34_Fnyes Vzmu Mjus 1.u. 45.
Lo Presti spring Karst water level
130
Karst water level (Fnyes) masl
500
median 1 = 476 mg/1
median 3 120
HCO3 mg/1
= 468 mg/1
110
450
100
400 90
1950 1958 1966 1974 1982 1990 1998 2006 2014
Date year
Figure23.9 Changes with time in the hydrogen-bicarbonate concentration based on some represent-
ative karst water wells, spring and seepages.
348 Karst without Boundaries
150 140
1.
Tata 28_Fnyes Vzmu Mjus 1.u. 43.
2.
Tata 34_Fnyes Vzmu Mjus 1.u. 45.
Lo Presti spring Karst water level
130
120
SO42 mg/1
median 2
= 59-67 mg/1
110
50
0 90
1950 1958 1966 1974 1982 1990 1998 2006 2014
Date year
Figure23.10 Changes with time in the sulphate concentration based on some representative karst
water wells, spring and seepages.
Reactivation of karst springs after regional mine dewatering 349
50 140
1.
Tata 28_Fnyes Vzmu Mjus 1.u. 43.
2.
Tata 34_Fnyes Vzmu Mjus 1.u. 45.
Lo Presti spring Karst water level
40 130
20 110
10 100
median 3 = 12 mg/1
0 90
1950 1958 1966 1974 1982 1990 1998 2006 2014
Date year
Figure23.11 Changes with time in the chloride concentration based on some representative karst
water wells, spring and seepages.
dropped to about 0.5 m3/s (Figure23.2). Figures 23.923.11 show the deepest point
in karst water level that was reached in 1990.
The remarkable changes in the concentration of hydrogen-bicarbonate around
198384 and of sulphate between 1977 and 19861987 are assumed to indicate the
hydrochemical response of the groundwater system due to large-scale aquifer dewa-
tering. As no sufficient data on hydrochemistry was available at the early stages of
the mining activity, nor on extraction rates, the lag between dewatering and chemical
response cannot be estimated confidently.
Recent hydrochemical data indicate that the concentration of the main water
components started to rebound towards their original values around the mid-2000s,
emphasised from 2010, presumably indicating the hydrochemical recovery of the
groundwater system. The comparison between abstraction rates and chemical data
suggests a delay of 1020 years between groundwater chemistry and groundwater flow
conditions.
Although further studies are required to support this conclusion, it can be assumed
that the hydrochemical changes were caused by the reversal of hydraulic gradients
and the subsequent changes in regional flow directions between the natural north
easterly flow and a modified southwesterly groundwater flow.
350 Karst without Boundaries
One of the major questions is whether the regenerating karst springs or new seepages
contain pure karst water, or are a mixture of karst and shallow groundwater, or are
seepages of shallow groundwater. Since very few samples could be analysed for iso-
topes, a conservative anion in the groundwater flow system, namely chloride, was
used to study mixing processes. Most of the dug wells in the city of Tata contain very
high nitrate concentrations, and sometimes ammonium, reflecting a strong anthro-
pogenic influence. Therefore, data from shallow boreholes were used (Figure23.8).
16 shallow boreholes were drilled in the drinking water protection area safety survey
(Csepregi, 2002). Some of them provided information on the connection between
shallow groundwater and karst water, others enabled an evaluation of the effects of
former pollution or potential pollution. The chemical composition of the water in
these shallow wells varies greatly, but for different reasons. The deeper wells show
connections either with the deeper groundwater, with the karst water, or one of the
wells in contact with the systematically regulated water level of Lake Tata. The vari-
able shallow groundwater composition is probably controlled by the fractures in the
area. Well TSZF-2/b shows a typical, shallow groundwater composition. This median
data for this well was used as an end member for shallow groundwater in the study
of potential mixing processes.
The concentration ranges with whiskers representing the 10 and 90% percentile
values of the main anions and their median values are shown in Figures23.1223.15.
In addition to the previously shown hydrogen-bicarbonate, sulphate and chloride
concentration time series of some representative karst water wells, springs and seep-
ages (Figures 23.923.11), these figures illustrate not just the typical concentration
ranges based on karst water wells and recent (Tata 28 and Tata 34 data of samples
collected in 2014 and 2015) karst water data (left side), but also the data of charac-
teristic shallow groundwater (TSZF-2/b), karst springs (Fnyes, Fnyes main, Pokol,
Lo Presti, Zsid iskola, Trkfrdo) and new seepages (Mjus 1.u. 43, Mjus 1.u.
45, Attila, Vadszbolt). It can be seen that the karst springs have similar concentra-
tion values as the karst water wells. While the CaMg-HCO3-water type of the new
seepages show a typical karst water composition, the higher Cl, NO3 concentra-
tions (first samples) at 45 Mjus 1 street also show an anthropogenic influence and
potential mixture with shallow groundwater. At the Attila and Vadszbolt seepages
a very small percentage of shallow groundwater mixing can also be seen, based on
the slightly higher hydrogen-bicarbonate and nitrate concentrations. The appear-
ance of nitrate in the karst springs and seepages show some local, minor anthropo-
genic contamination of the shallow aquifer which through mixing with karst water
results in elevated concentrations of these parameters in certain springs. The mixing
between karst water and shallow groundwater primarily takes place along deep
fractures where the Mesozoic carbonate reservoir discharges to the surface.
800
750
700
HCO3 mg/1
600
502
482
482
478
477
476
471
470
468
463
459
459
459
458
458
458
457
500
452
400
347
300
Trkfrdo
TSZF-2/b
Vadszbolt
Attila
Tata 34 (I.)
Fnyes main
Zsid iskola
Tata 28 (I.)
Polkol
Tata 26
Karst recent
Lo Presti
Fnyes
Mjus 1.u. 45
Tata 28 (II.)
Tata 34 (II.)
Tata 41
Mjus 1. u. 43.
Tata 27A
Bds csorg
The tritium content of the sampled karst waters is below the detection limit
(0.059 Bq/l), which suggests that there is no precipitation component younger than
50 years in the karst waters (Figure23.16). This also means that shallow ground
water does not mix with karst waters at the surveyed sites.
Based on previous studies (Dek, 2006) groundwater in Hungarian aquifers with
18O values higher than 10 are generally considered to be of Holocene origin.
However, where mixing has occurred, especially with deep formation waters, higher
values are also recorded. The 18O data of the karst wells cluster very close to each
other with values of 10.9, while the D data range between 75.1 and 76.2
respectively (Figure23.17). The data (6.6 18O and 52.2 D) of the Bds csorg
spring which is shifted from the Carpathian Basin Meteoric Water Line (CBMWL)
D = 7.8*18O + 6 defined by Dek (1995) shows the effect of evaporation (Clark &
352 Karst without Boundaries
400
355
300
247
SO42 mg/1
200
79
67
100
64
63
62
61
59
56
58
55
52
50
51
51
46
10
10
10
0
TSZF-2/b
Bds csorg
Attila
Tata 34 (II.)
Mjus 1. u. 43.
Vadszbolt
Fnyes main
Tata 41
Tata 28 (II.)
Tata 27A
Mjus 1.u. 45
Karst recent
Zsid iskola
Trkfrdo
Lo Presti
Fnyes
Polkol
Tata 28 (I.)
Tata 26
Tata 34 (I.)
Figure23.13 Sulphate concentration ranges with whiskers representing the 10 and 90% percentile
values and their median values (left: karst water wells; right: shallow groundwater
(TSZF-2/b), karst springs and new seepages).
Fritz, 1997). The CBMWL is very close to the Global Meteoric Water Line defined
by Craig (1961).
Age calculations based on Carbon-14 measurements were done by the application
of a Carbon-13 correction (Pearson, 1965). Carbon-14 data vary between 8.59.7%
pmC, while 13C data vary between 8 and 9. The radiocarbon ages calculated by the
widely used Carbon-13 correction method vary between 11800 years and 13100 years.
The isotope compositions of the sampled karst waters clearly show Pleistocene
recharge except for well K-28. It can, therefore, be assumed that the karst waters in
the Tata region are older than 10000 years. Importantly, despite a rise in the karst
water level, no sign of recent (young) infiltration could be detected in the karst water.
Bds csorg spring and a dug well at the Kismos stream have a significant or wholly
recent infiltration origin. Based on tritium data, mixing with karst water cannot be
Reactivation of karst springs after regional mine dewatering 353
100
88
90
72
80
70
54
60
CI mg/1
50
40
22
22
30
21
21
20
20
19
18
18
17
16
15
15
14
13
12
20
9
10
0
Attila
Polkol
Fnyes
Tata 41
Tata 26
TSZF-2/b
Tata 27A
Lo Presti
Tata 34 (I.)
Vadszbolt
Tata 28 (I.)
Trkfrdo
Mjus 1.u. 45
Tata 34 (II.)
Tata 28 (II.)
Mjus 1.u. 43.
Zsid iskola
Fnyes main
Bds csorg
Karst recent
Figure23.14 Chloride concentration ranges with whiskers representing the 10 and 90% percentile
values and their median values (left: karst water wells; right: shallow groundwater
(TSZF-2/b), karst springs and new seepages).
completely excluded, but neither the main nor the trace element data support the pos-
sibility of mixing. The high chloride concentration (Figure23.14) shows mixing with
shallow groundwater. More information on mixing could be gained by using D18O
data. Figure23.18 shows that with an increase of the recent infiltration component,
shown by an increase in the tritium content, there is a significant increase in the chlo-
ride concentration which suggests mixing with shallow groundwater at these sites.
23.9 CONCLUSIONS
105,0
110
100
90
NO3 mg/1
10
5,2
3,9
2,3
2,2
0,0001
0,0001
0,0001
1,1
0,8
0,7
0,7
0,6
0,5
0,5
0,5
0,0001
0,0001
0,2
0,3
Attila
Polkol
Fnyes
Tata 41
Tata 26
Lo Presti
Tata 27A
TSZF-2/b
Mjus 1.u. 45
Vadszbolt
Tata 28 (I.)
Tata 34 (I.)
Tata 28 (II.)
Tata 34 (II.)
Zsid iskola
Trkfrdo
Fnyes main
Bds csorg
Karst recent
Figure23.15 Nitrate concentration ranges with whiskers representing the 10 and 90% percentile
values and their median values (left: karst water wells; right: shallow groundwater
(TSZF-2/b), karst springs and new seepages).
Spring locations are aligned with deep tectonic structures both in uncovered and
confined karst areas. This indicates that concentrated groundwater flow takes place
along tectonic structures and the confining marls do not prevent groundwater dis-
charge to the surface.
The comparison between time series of monitoring wells screened in the shallow
aquifer and those installed in the deep karst aquifer suggest that there is no hydraulic
connection between shallow and karst aquifers.
The prediction of karst water levels based on physical curve fitting suggests that
equilibrium level of the karst water table will be reached around 2018 at approxi-
mately 140 m asl. While most springs have already reactivated in recent years, the Lo
Presti, Kismos, Pokol and Tkr springs are expected to reactivate until 2018. The
water level is predicted to rise by a further 4 m.
Reactivation of karst springs after regional mine dewatering 355
7 Evaporation effect 5.
1.
9,6
8
2.
18OVSMOW
C pmC %
9 9,2
14
10 1. Majus 1.u. 43.
2. Fenyes spring
3. Trkfrdo 8,8
1. 4. Tata 28_Fenyes Vzmu
1.
2. 4. 5. Bds csorg
11 3.
Pleistocene waters
18O data
3. 14
C data
12 8,4
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
3
H Bq/l
The karst waters in the Tata region are CaMg-HCO3 type, indicating that they
come from a dolomitic aquifer. While the karst waters of the region show a uniform
composition, the shallow groundwater shows a variable composition. The chemical
composition of the surveyed wells, the Fnyes, Lo Presti, Trkfrdo, Pokol springs,
and the new Attila seepage and seepages at 43 and 45 Mjus 1 street show karstic
water composition. The chemical composition of Kastlykert spring and the dug well
at Kismos stream are similar, and show the effect of mixing with shallow ground
water and potentially also local pollution. Kismos stream does not show any connec-
tion with the karst water and is badly polluted.
Geochemical data indicate significant changes in karst water chemistry in response
to groundwater depressurisation and later recovery. While bicarbonate concentra-
tions decreased, sulphate concentrations increased during the mine dewatering oper-
ations. Recent hydrochemical data indicates that the concentration of the main water
components started to recover towards their original values around the mid-2000s,
increasingly since 2010. This presumably indicates the geochemical recovery of the
groundwater system. The available data suggest a delay of approximately 1020 years
between the changes in dewatering rates and the subsequent hydrochemical reactions.
The hydrochemical changes were caused by the reversal of hydraulic gradients and
regional flow directions between the natural north easterly flow and a modified south
westerly groundwater flow direction.
50
Evaporation effect
60
Holocene infiltration
DVSMOW
70
Karst water
Majus 1.u. 43, Fenyes spring,
80 Trk frdo , Tata 28_Fenyes Vzmu 1.
GMWL
CBMWL
Recent infiltration
Bds csorgo
90
12 10 8 6
18OVSMOW
100
9.
80
60
Cl mg/l
8.
40
1. Fenyes spring
2. Majus 1.u. 43.
3. Trkfrdo
7. 4. Lo Presti
20 1. 6. 5. Tata 28_Fenyes Vzmu 1.
2. 6. Vadaszbolt
5.
3. 7. Kismoso stream (seepage)
4. 8. Kismoso well
9. Bds csorgo
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
3
H Bq/l
Chloride as a conservative anion was used to study the mixing processes. Most
springs have similar concentration values as karst water wells. While new seep-
ages show a typical karst water composition, the higher Cl, NO3 concentrations at
45 Mjus 1 street also show an anthropogenic influence and potential mixing with
shallow groundwater. At the Attila and Vadszbolt seepages a very small percentage
of shallow groundwater mixing can also be seen.
The tritium content of karst waters is below detection limit, which means there
is no precipitation component younger than 50 years in the karst waters. The isotope
compositions of the sampled karst waters clearly show Pleistocene recharge except
for well K-28. Based on this it can be assumed that the karst waters in the Tata region
are older than 10000 years. Importantly, despite recent rising water levels no signs
of recent infiltration could be detected in the karst water discharges. Bds csorg
spring and a dug well at the Kismos stream have a significant or wholly recent
infiltration origin.
ACKNOWLEDGMENTS
We would like to thank the City Council of Tata for providing support and making
their data available for this research. We also acknowledge the North Transdanubian
Water Directorate, the szakdunntli Vzmu Zrt., Mrton Maller, Gza Hajnal
and Andrs Csepregi who provided data for this study, and Fnyes Spa for providing
local knowledge. We would like to thank Zsolt Galgczy and Andrs Rehk for their
field help.
REFERENCES
Clark I., Fritz P. (1997) Environmental isotopes in Hydrogeology. Lewis Publishers, USA.
Cooper H.H., Jacob C.E. (1946) A generalized graphical method for evaluating formation
constants and summarizing well field history, American Geophysical Union Transactions 27,
526534.
Craig H. (1961) Isotopic Variations in Meteoric Waters. Science, 133, pp. 17021703.
Csepregi A. (2002) A tatai vzbzis biztonsgba helyezsi terve. Report of Csepregi Ltd.
Csepregi A. (2007) A karsztvztermels hatsa a Dunntli-kzphegysg vzhztartsra. In:
Bnyszati karsztvzszint sllyeszts a Dunntli-kzphegysgben. (Ed.: Alfldi, L, Kapolyi
L.) pp. 77106.
Dek J. (1995): A felszn alatti vizek utnptldsnak meghatrozsa izotpos mdszerekkel
az Alfldn. VITUKI report. Budapest.
Dek J. (2006): A Duna-Tisza kze rtegvz ramlsi rendszernek izotp-hidrolgiai vizs-
glata. PhD thesis. Budapest.
Fogarasi S. (2001) Visszatrnek-e a tatai forrsok? Fldrajzi Konferencia, Szeged. Report. 15.p.
Horusitzky, H. (1923) Tata s Tvros hvforrsainak hidrogeolgija s kzgazdasgijvoje
A Magyar Kirlyi Fldtani Intzet vknyve XXV. ktet 3. fzet, Budapest. pp. 3783.
Maller M., Hajnal G. (2013) A tatai forrsok hidrogeolgiai vizsglata. Mrnkgeolgia-
Kozetmechanika (Ed.: Trk ., Grg P. & Vsrhelyi B.), pp. 718.
Pearson F. J., Jr. (1965) Use of 13C/12C ratios to correct radiocarbon ages of materials initially
diluted by limestones, In: Chatters, R M and Olson, E A, (Eds.), Intern. conf. on 14C and
tritium dating, 6th, Proc: Clearinghouse Fed Sci Tech Inf, NBS, Washington, DC, p 357366.
358 Karst without Boundaries
Theis C.V. (1935) The relation between the lowering of the piezometric surface and the rate
and duration of discharge of a well using groundwater storage, American Geophysical Union
Transactions 16, 519524.
VITUKI Rt. (2000) Karsztvzvdelem a Kzp-Dunntli Rgiban, Budapest. Report.
(T.: 721/1/4846-1)
Geographic names (Regional)
23
Editors:
Karstified rocks, of various lithologies, cover more than 10% of the land surface of
Neno Kukuric
Neven Kreic
Zoran Stevanovic
Editors:
an informa business